Hamid N. Alsadi
Seismic
Hydrocarbon
Exploration
2D and 3D Techniques
Advances in Oil and Gas Exploration &
Production
Series editor
Rudy Swennen, Department of Earth and Environmental Sciences,
K.U. Leuven, Heverlee, Belgium
The book series Advances in Oil and Gas Exploration & Production pub
lishes scientic monographs on a broad range of topics concerning geo
physical and geological research on conventional and unconventional oil and
gas systems, and approaching those topics from both an exploration and a
production standpoint. The series is intended to form a diverse library of
reference works by describing the current state of research on selected
themes, such as certain techniques used in the petroleum geoscience business
or regional aspects. All books in the series are written and edited by leading
experts actively engaged in the respective eld.
The Advances in Oil and Gas Exploration & Production series includes
both single and multiauthored books, as well as edited volumes. The Series
Editor, Dr. Rudy Swennen (KU Leuven, Belgium), is currently accepting
proposals and a proposal form can be obtained from our representative at
Springer, Dr. Alexis Vizcaino (Alexis.Vizcaino@springer.com).
Seismic Hydrocarbon
Exploration
2D and 3D Techniques
123
Hamid N. Alsadi
Data Processing Section
Ministry of Oil
Baghdad
Iraq
The favorable reception of the rst edition of this book (seismic Exploration,
published by Birkhauser Verlag in 1980) stimulated my belief in the need of
an updated book that includes the advances in the techniques which have
taken place during the past three decades. In preparing the present updated
volume, I have taken into consideration the remarks and suggestions of the
users of the 1980 edition from both of the academic and industrial work
domains.
Since 1980, when the rst edition of this book was published, great
developments in the seismic exploration technology have taken place. These
developments have occurred in all of the three exploration phases: acquisi
tion, processing, and interpretation techniques. The most prominent advances
which have taken place in these years are the widespread implementations
of the 3D surveying, prestack migration, and growing interpretation tech
niques in both structural and stratigraphic exploration. As it is familiar with
the exploration geophysicists, this subject (seismic exploration) is fully dealt
with in many original and authentic internationally known text books. In this
publication, no new subjects added to those found in the other standard books
which are well known in the geophysical library. In fact, these and other
related scientic papers and research reports formed solid references for the
present book. There are, however, differences in the design and presentation
approach.
In its design, the book is intended to be a comprehensive treatise of the
seismic exploration tool, addressing audiences in both of the academic and
industrial establishments. It is made up of 12 chapters covering the basic
aspects of the seismic reflection exploration subject, starting with the basic
theory, followed by the applied data acquisition technology, and ending with
the processing and interpretation. In presenting the subject matter, emphasis
is made on the practical aspects of the subject, using clear and simplied
presentation, avoiding excessive descriptions and unnecessary lengthy
comments. Numerous illustration gures (>390 gures) have been used
throughout the book to aid in clarifying the concepts and procedures involved
in any standard seismic exploration survey. In this way, the book can be
considered as a very useful introductory teaching manual for university
students taking seismic reflection exploration as part of a postgraduate
course.
v
vi Preface
The chapters of the book are sequenced in the order of the activities
normally executed in a standard seismic exploration survey: eld acquisition,
processing, and interpretation. Chapter 1 is an introductory chapter in which
a brief historical note and short review of the geophysical exploration
methods are given. This is followed by four chapters covering the theoretical
aspect of the subject including basic principles and denitions of seismic
waves, with a special chapter assigned for the seismic wave propagation
velocity. The propagation phenomena, reflection, diffraction, transmission,
and refraction, are dealt with in Chaps. 4 and 5. The following two chapters
are devoted to the two main tools applied in seismic exploration, namely the
2D and 3D surveying techniques.
Due to its important role in understanding of the processing steps applied in
seismic data processing, a chapter (Chap. 8) is assigned solely for the seismic
signal. This chapter is structured on the theme of considering the seismic
reflection wavelet as a propagating signal in the same way as the electro
magnetic signal is treated by the communication theory applied in electro
magnetic wave propagation. Including a chapter on seismic signals, preceding
the processing chapters, is a feature by which this book has deviated from
other conventional publications. Data processing is presented under two
headings: processing tools (Chap. 9) and the normally applied processing
sequence (Chap. 10). Chapter 11 covers some specialized seismic exploration
tools sometimes used in support of the conventional seismic reflection and
refraction surveying. The book is concluded with Chap. 12 on interpretation.
I would like to express my gratitude to my wife Asira and my sons (Eng.
Majid, Ph.D., Eng. Muhannad, M.Sc., Thurayah, B.Sc., and Eng. Mahir,
B.Sc.), for their continuous support and help throughout the past three years.
My work in the writing of the book has incurred an additional burden to the
family especially during the abnormally difcult times, which our country
has experienced in the past twelve years.
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Historical Review . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 Elasticity and Seismic Waves (Bth 1979) . . . 1
1.1.2 Earthquake Seismology (Richter 1958). . . . . . 2
1.1.3 Exploration Seismology
(Telford et al. 1990) . . . . . . . . . . . . . . . . . . 2
1.1.4 Summary of Exploration Seismology History
(Sheriff and Geldart 1995) . . . . . . . . . . . . . . 3
1.2 The Geophysical Exploration Project . . . . . . . . . . . . . 3
1.3 Geophysical Exploration Methods . . . . . . . . . . . . . . . 4
1.3.1 The Seismic Method . . . . . . . . . . . . . . . . . . 4
1.3.2 The Gravity Method . . . . . . . . . . . . . . . . . . 4
1.3.3 The Magnetic Method . . . . . . . . . . . . . . . . . 6
1.3.4 The Electical Method. . . . . . . . . . . . . . . . . . 7
1.3.5 The Radioactivity Method . . . . . . . . . . . . . . 9
1.4 Oil Well Drilling. . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4.1 Drilling of the ExplorationWell . . . . . . . . . . 10
1.4.2 The OilWell Rotary Drilling . . . . . . . . . . . . 10
1.4.3 The DrillHole and Well Casing . . . . . . . . . . 11
1.5 Well Geophysical Logging . . . . . . . . . . . . . . . . . . . . 11
1.5.1 Electrical Logging . . . . . . . . . . . . . . . . . . . . 12
1.5.2 Radioactivity Logging . . . . . . . . . . . . . . . . . 14
1.5.3 Acoustic Logging . . . . . . . . . . . . . . . . . . . . 16
1.5.4 Log Interpretation . . . . . . . . . . . . . . . . . . . . 19
1.6 Latest Developments in Well Logging . . . . . . . . . . . . 19
1.6.1 Logging While Drilling Technique . . . . . . . . 20
1.6.2 Borehole Imaging Tools. . . . . . . . . . . . . . . . 20
1.7 Well Completion . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2 Seismic Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.1 The Fundamental Conditions . . . . . . . . . . . . . . . . . . 23
2.2 Theory of Elasticity . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.1 Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.2 Strain . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.2.3 Common Types of Strain . . . . . . . . . . . . . . . 26
2.2.4 The VolumeChanging Strain . . . . . . . . . . . . 27
2.2.5 The ShapeChanging Strain . . . . . . . . . . . . . 28
2.2.6 The Cubical Dilatation . . . . . . . . . . . . . . . . . 28
vii
viii Contents
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Introduction
1
waves, named after its discoverer, the English In recent times, we may add AgadirMorocco
man A. E. H. Love. in 1960, KobeJapan in 1995, FukushimaJapan
in 2011, and KatimandoNipal in 2015.
seismic
reflection depth
raypath
reflector
seismic
refraction depth
raypath
reflector
Fig. 1.2 Essence of the two seismic techniques; reflection (a) and refraction (b)
change (called the gravity anomaly) is greater The measuring instruments, normally used in
over the anticline axis falling to lower values gravity surveying, are called the gravity meters
over the flanks. A structural feature showing (or gravimeters). These are designed to measure
relative densitydeciency (such as a salt dome), gravity variations rather than absolute gravity
gives an inverted bellshaped anomaly normally values. These instruments are capable to measure
referred to as (negative anomaly) which is of gravity changes to less than a tenth of a milligal.
opposite shapechanges to that exhibited by the The gravity unit (the gal) is dened to be equal to
relatively denserrock geological anomaly (posi 1 cm/s2. Thus the gravimeter is capable of mea
tive anomaly). The principle is shown in Fig. 1.3. suring gravity variations to within about
0 0
Sea level
geological anomaly
1 2 1 2
Fig. 1.3 Dependence of the algebraic sign of surplus where body density (q2) is greater than that of the
gravityanomaly on whether the contrast is surplus or host medium (density, q1), and vice versa, given (q2 > q1)
deffeciency. Positive sign is given to case of density
6 1 Introduction
L L L
Fig. 1.5 Principle of the Wenner conguration used in electrical resistivity surveying. I and V denote electric current
and voltage respectively. L is the distance between electrodes
accumulations.
NaturalSource Electrical Methods
time, t There is another less commonly applied electrical
t0
group of methods which are dependent on mea
Fig. 1.6 The procedure used in the IP method. Voltage surements of the earth natural electric eld. Self
drop after the applied electric current is switched off at Potential (SP), Telluric Current and Magneto
time (t0)
Telluric methods are typical examples of such
techniques. These methods are used to determine
back to normal state after the applied electric largescale crustal structural variations as sedi
current is suddenly stopped at a certain instant. The mentary basins and regionalscale geological
process is sketched in Fig. 1.6. variations.
The method is used mainly for purposes of (i) Self Potential (SP) Surveying
metallic minerals and ground water exploration.
Mineralized zones found in the upper part of the
For the eld surveying, the same electrode set up
Earth crust and some ore bodies found at shallow
for resistivity is also used. Thus using four depths, develop their own natural electric elds.
electrodes (called dipoledipole array) is used in
In particular, a metallic sulde body which has
IP surveying. The polar dipole system is pre
part of it above the water table and the rest of it, is
ferred where the voltage electrode spacing is below it. With this condition, such a phenomenon
decreased to minimize the effect of wirecurrents.
is created. The body acts as a natural battery
(iii) Equipotential and Electromagnetic where ions, due to differential oxidation, move
(EM) Surveying from one part of the body to the other. The
Equipotential surveing uses the electric potential electrical currents are formed as a result of
eld generated by a xed electriccurrent electrode chemical reaction that takes place with the aid of
while the electric potential is mapped by a moving the electrolytes present in the host medium.
electrode. Instead of using direct current The potential difference measured on the sur
(DC) electrodes, another method, the Electro face will map the potential anomaly caused by the
magnetic (EM) method, uses an electric alternating body. The behavior of the electric potential due to
current (AC) of few hundredtofew kilohertz, presence of a sulde body is shown in Fig. 1.7.
induced into the ground by a sourcecoil and
received by another coil, the receivercoil. The electrical
transmitter coil induces an alternating (primary) potential
magnetic eld, while the receiver coil senses the +
groundgenerated alternating (secondary) mag 0
netic eld combined with the primary. The com
bined magnetic elds induce AC in the secondary 
coil. This current is then measured and converted ground level
into the combined magnetic eld at the location of
oxidization
the receiver. From comparison of the combined region
magnetic eld with the primary magnetic eld + water table
S
(which is known), the contribution of the anoma 
lous body is determined and then interpreted into
current
geological information. flow lines
In addition to ground surveying, there is the
airborne EM surveying where the transmitter coil Fig. 1.7 Sulde body (S) acting as a natural battery,
is xed to the aircraft and receiver is mounted in causing abnormality in the generated electrical potential
1.3 Geophysical Exploration Methods 9
The SP method is typically used to map the summarized in the following table (Maton et al.
electrochemical potential variations, generated 1995, p. 269).
by buried sulde bodies.
Radiation Radiation Energy
(ii) Telluric Currents Surveying type characteristics characteristics
The term (telluric currents) is used for the natural
Alpha Positivelycharged Can be stopped
electric currents that flow horizontally in the particles protons by a sheet of
upper part of the Earth crust. Variation of the paper
currentdensity, over the earth surface is gov Beta Negativelycharged Can pass
erned by the rockresistivity changes. Thus, if a particles electrons through as
salt dome, for example, is found imbedded much as 3 mm
of Aluminum
within relatively high conductivityformations,
the lines of currents flow will bypass the salt Gamma Highfrequency Can pass
rays electromagnetic through several
body causing distortions in the potential gradient waves centimeters of
in the overlying surface cover. From the detected Lead
anomalous potential gradient, the geological
changes causing the changes in the potential All of the three types of radiation have the
gradient, can be determined. ionizing ability when colliding with atoms. This
(iii) MagnetoTelluric Currents Surveying effect is the basis of radioactivity detection.
Another related method, the socalled A detection instrument (the Geiger counter)
(magnetotelluric method), involves simultane was designed by Hans Geiger in 1928. It consists
ous measurements of both voltage created by the of a tube lled with an inert gas (helium or
telluric currents and the magnetic eld induced argon) at a reduced pressure. When radiation
by these currents. Plots of the alternating voltage enters the tube, it removes electrons from the
and that of the associated magnetic eld (as atoms of the gas, making them positively
function of frequency) can give information on charged ions. The electrons move to wire in the
the resistivity as function of depth. tube, setting up an electric current which is
amplied and fed into the recording device. The
(iv) AudioFrequency Magnetic Surveying current produces clicking sound that pulsates at a
There is still another closely related method rate which is depending on the radiation strength.
which uses the audiofrequency variations in the A counting unit (counter), attached to the wire,
Earth electric eld, to study the earth resistivity measures the created current and gives a pro
variations. This method is called Audio portional radioactivityreading.
Frequency Magnetic method, termed (AFMAG)
method.
Fig. 1.9 A simplied sketch showing the main parts of a 1.5 Well Geophysical Logging
rotary drilling rig
S
r1
r2
normalsonde lateralsonde
configuration configuratio
sonde
SPLog
only in an uncased holes lled with conductive be drawn for sandstone SPvalue. The inflection
mud. The measuring sonde is of simple congu points in the SPlog indicate formation
ration. It consists of only two electrodes: one is boundaries.
lowered into the well by an insulated cable and
the other is xed at the ground surface
(Fig. 1.15). 1.5.2 Radioactivity Logging
This method, which needs no articial current,
can detect natural potential differences which Radioactivity is the phenomenon of emission of
develop at formation boundaries. Compared with particles and photons of electromagnetic energy
resistivity logs, the SP logs give more sharp from an atom. This radiation process occurs
changes and hence more accurate formation either naturally from unstable nuclii or induced
boundaries. Like lateral logging mentioned by bombarding of stable nuclii by photons or
above, there is a variation made on induction atom particles. Examples of such radioactive
logging to give focused current radiation for elements are Uranium, Thorium Rubidium and
getting sharper boundary indication. Potassium 40 which is most commonly found in
Interpretation of the SP logs depends on the shale and clay and less in sandstone and
manner of the electrical potential variation. As a limestone.
general rule, changing of the log values towards There are three types of radiation: Alfa parti
positive potentials is considered as an indicative cles formed of charged helium nuclii, Beta par
of impermeable rocks as shale or tight limestone ticles formed of high speed electrons, and
or tight sandstones. When the change is towards Gamma rays of electromagnetic waveenergy.
negative potential, it is interpreted as being due Out of these, only gamma ray is used in well
to porous sandstones or porous limestone. In this radioactivity logging. The detection instrument is
way, the SP logs give useful indications on a Geiger or more usually is the scintillation
lithology, water salinity and help in determina counter which consists of a special crystal (like
tion of formation boundaries. In particular, it sodium iodide), which emits flashes of light as
shows the boundaries of formations in sandshale they absorb gammaray photons, hence the name
sequences. These are determined with the help of (scintillation counter). A photoelectric tube con
drawing lines (sand and shale baselines) in the verts these flashes into electric currents which are
produced SP log (shown in Fig. 1.15). It is displayed in the form of a continuous chart (the
commonly observed that shale beds give the radioactivity log).
same level of SPreadings allowing for drawing a Radioactivity logs provide important infor
straight line indicating shale SPvalue. This is mation on rock lithological types, especially on
called (the shale line). Similarly a (sand line) can those containing certain concentrations of
1.5 Well Geophysical Logging 15
shale
S S
limestone
R R
R R sandstone
S
shale
about 3 ft from the nearest receiver. To correct Sonic logs provide the interval transit time
for tilting and holeirregularities effects, a dual which represent traveltime (usually in
source sonde is used, making what is called a microsecond units) of the Pwave in covering a
boreholecompensated sonde (Fig. 1.18). distance (usually 1 ft). This parameter (the
The electronic structure of the sonde is interval transit time) is useful in computing the
designed in such a way, that the output is made porosity () by using the following relationship:
to be the difference in the traveltimes to the two
receivers. The time difference, measured in Dtl Dtm =Dtf Dtm
timeunits per 1ft, called (interval transit time),
is plotted (normally in micro seconds) against where, (Dtm), (Dtf), and (Dtl), represent transit
depth to give the continuous wiggly curve known times of matrix material, of pore fluid, and log
as the (sonic log). In the compensated sonic readtransit time respectively.
logging, seismic pulses are emitted alternately
(ii) Well Velocity Surveying
from the two sources and the transit times from
A hydrophonetype detector is lowered down the
the two oppositely traveling refracted Pwaves
well which is lled with the drilling fluid. The
are averaged electronically. The output (transit
travel time of the seismic wave generated by a
time) is plotted against depth, giving the
surfaceplaced shot and received by a detector
boreholecompensated (BHC) sonic log
placed at a formation boundary is recorded. The
(Fig. 1.19).
full log is obtained by repeating the recording at
The borehole compensated sonde (BHC)
each boundary traversed by the well. From the
gives an average interval transit time which is
measured travel time of direct arrivals, the av
plotted on a paper strip. The produced log in this
erage velocity and interval velocity are plotted as
case is normally referred to as BHC sonic log
function of welldepth. The velocity survey is
which is used to identify lithologies, determine
also called (checkshot survey). Principle of the
formation boundaries, and in computing syn
velocity survey setup and velocitydepth plot is
thetic seismograms. The interval transit time can
shown in Fig. 1.20.
be integrated down the well to give the total
The obtained velocity information is used to
travel time. This type of logging can only be run
calibrate the sonic log and check the soniclog
in an open (uncased) hole.
18 1 Introduction
seismic
detector
depth
integrated time (hence the name, checkshot normally applied to refer to the direct and
surveying). reflected waves respectively. The ray paths of the
VSP shooting and the corresponding traveltime
(iii) Vertical Seismic Proling (VSP) plot of the recorded waves are schematically
Basically the measurement setup is the same as shown in Fig. 1.21.
the checkshot recording system. The difference A seismic pulse is generated on the surface
is in the recording duration time which is here from dynamite explosion or from an air gun
extended to allow recording reflected waves as submerged in a water lled hole. This is repeated
well as the direct arrivals. At each detector at all of the hydrophone positions, normally
stoplocation (normally at 25 m spacing), the positioned at 25 mspacing down the well.
recorded seismic trace is allowed to include With an appropriate processing sequence, the
events from up going reflected waves in addition nal corrected VSP section is produced. Process
to the rst arrival (direct wave) event. The terms ing includes, data editing, correction to vertical
(downgoing wave) and (upgoing wave) are time and velocity ltering for separating unwanted
R1
down tube upgoing wave
R2 going wave
wave
R3
Raypaths of Traveltime plot
direct and reflected waves of the tube wave,
from a threereflector model downgoing & upgoing waves
1.5 Well Geophysical Logging 19
events. It is evident from the VSP section that Drilling information (drilling rate, mud den
downward events (primary and multiples) increase sity variation)
in time with depth, and the upward events (primary Caliper and dipmeter logs.
and multiples) decrease in time with increasing
The main information which can be obtained
depth. By arranging the produced seismic traces
from well logs can be summarized as in the
(25 m spaced down the hole) a seismic VSP sec
following table
tion is obtained. Each seismic trace of a VSP
section contains events from downgoing waves
Log Type Application Comment
(direct and multiples arrival) and from upgoing
Electrical Fluidtype Logging is
waves (reflections and multiples). Logs identication done in
The main application of VSP is providing Resistivity Porosity uncased
seismic section of reflectors which have not been Induction evaluation wells
reached by the drilling. In fact the VSP data are Spontaneous Boundary
Potential determination
equivalent to both the checkshot data and the Shale and sand
sonicderived synthetic seismogram. lines
(iv) Other Logging Techniques Radioactivity Shale/sandstone Logging is
Logs Formation done in
Some logs are used to give geometrical infor
Natural density cased wells
mation on the deviation from the vertical gammaray Porosity
(holedrift angle) as well as the azimuth of the Gammaray Fluid type
deviation. Other logs are made to give the bed density
Neutron
ding dip (dipmeter logs) and measurements of
gammaray
the wellhole diameter (caliper logs).
Acoustic Lithology types Logging is
On more limited scale of application, other Logs Formation done in
types of well logging have been used. Specially Sonic logs Boundaries cased and in
modied boreholegravimeters, magnetometers, Well Synthetic uncased
and thermometers are examples of such tools. velocity seismograms wells
VSP Seismicvelocity
Geothermal prospecting is applied to detect functions
geological features which affect heat flow such as Reflection
shallow salt domes, faults and dykes. Also identication
ground water is investigated by this method.
1.6.1 Logging While Drilling runningin and cementing the casing is consid
Technique ered as integral part of the completion process.
To prevent caving in of the well walls, drilling
In late 1980s a new logging procedure, logging is stopped at various depths. and a steel pipe
while drilling (LWD), was introduced. This (casing) is lowered into the well and xed in
technique is similar to the conventional wireline position by pumping cement mixture in the space
logging except that it (the LWD) uses sensors between the casing pipes and the well walls
which are incorporated in the drilling bit (cementing process). According to the planned
assembly, and thus the measurements are made drilling program, drilling after casing being
during the drilling process. This measurement cemented is continued with a different (usually
procedure coupled with specialized computer smaller) bitsize. At this stage, the drilled hole is
software provides fast geological information as lined by the appropriate casing which is likewise
lithology, porosity, fluid contents, and drillhole cemented in. The drillcasingandcementing
direction. phases are repeated until the nal planned total
depth is reached. The end result of the casing
operations is a set of concentric pipes each of
1.6.2 Borehole Imaging Tools which is ending at the earth surface. When the
planned total depth is reached, all used drill pipes
Borehole imaging is a logging technique that can including the attached drilling bit, are withdrawn.
provide smallscale images of the borehole wall The next operation is perforation of the parts of
and the penetrated rock lithology. One of these the casing facing the oil bearing formations. The
methods (based on electrical microimaging soproduced holes are made to allow the oil to
processes) was developed by Schlumberger
under the tradename; formation microimaging
(FMI). In 1986, Schlumberger developed a
dipmeter called the formation microscanner christmas
(FMS). Later versions of this logging tool were tree
called formation microimager (FMI) which is
based on microresistivity measurements.
The operation of the FMI tool is based on
recording changes (and not absolute values) in
electrical resistivity of rock formations. It pro
vides, with high resolution power, identication
20inch casing
of sedimentary characteristics and fracturing
picture in the penetrated rock column, in addition
to measuring formation dip. 13 inch casing
seep inside the central production tubing of the equipment in addition to the mechanical system
well. that provide the blowout prevention (BOP) is
The last stage in preparing the well for the called the well head which is also equipped with
production process is xing of a group of flanges an assembly of valves that controls flow, and any
which provide a sealing structure connecting other possible interventions. This assembly is
together all the surface ends of the casing tubes. normally referred to as the Christmas tree
This part of the well which contains pressure (Fig. 1.22).
Seismic Waves
2
2.1 The Fundamental Conditions motion of elastic waves are generated and can be
recorded by the appropriate detection instru
An application of an external force, on part of a ments. These are normally referred to as the
medium (elastic medium), leads to creation of (seismic waves).
internal opposing forces which intend to resist
the deformations caused by that external force.
Typical forms of the resulting deformations are 2.2 Theory of Elasticity
changes in volume and/or in shape which are
created at the affected location. In consequence, As it is stated above, the fundamental conditions
the medium will return to its original condition for a seismic eld to be created is that the med
after the external force is removed. This property ium must possess the elasticity property. Two
of resisting of changes in volume and in shape main concepts are governing the propagation of
and return to original conditions after removal of seismic waves in an elastic medium: the (stress)
the external force is called (elasticity). Provided and the (strain). Stress represents the external
that the changes are small, rock media in nature force applied to the elastic medium, and strain is
are considered to be perfectly elastic in nature. the resulting changes in volume and in shape.
As a result of the elasticity property of media, The relation between stress and strain, for a
the changes (volume and shape changes) oscil particular medium (perfectly elastic medium),
late about their neutral positions and, at the same gives evaluation expressions for the elasticity
time, propagate away from the energy property of that medium. The stressstrain pro
sourcelocation. Energy transfer in this manner portionality constants are the elastic coefcients
(motion that leaves out no permanent distortions) which serve as measures of the elasticity of a
is commonly referred to as (wave motion). particular medium.
The fundamental condition for the creation The principal types of changes experienced by
and propagation of seismic waves (seismic eld) a medium due to passage of a seismic wave are
is a source of mechanical energy of impulsive redistribution of the internal forces (stress
type which is initiated within an elastic medium. changes) and modication of the volume and
The energy source may be natural (as in geometrical shape (strain changes). The theory of
earthquakegenerated waves) or articial (as in elasticity deals with analysis of these principal
ring of a dynamite charge). In both cases wave effects and the related physical changes.
T
Tz
Ty
Tx
A
Fig. 2.1 Stress (T) and its components (Tx ,Ty, Tz) acting on the elementary area (DA)
2.2 Theory of Elasticity 25
C
y
z
The three stress components (Tx , Ty , Tz) per each of the three
areas (A, B, C)
Tyz B
A
T
Txz Tzz
T T
Tyx
Tyy
C
Txx Tzx
Txy Tzy
Fig. 2.2 The stress components. The mutually perpendicular planes (A, B, C) making up the areaset and the nine
components of the involved stress tensor
meter (N/m2) which is called Pascal, where one In general, when a body is subjected to elastic
Pascal is equal to 1 N/m2. stress, both of its size and shape will change. As
it is mentioned above, the resulting changes
represent elastic strains when each point of the
2.2.2 Strain stressed body experiences a displacement of its
own which is different from the displacements
In reference to Fig. 2.3, let us consider the two experienced by the other points of the body. This
points (P1 and P2) located within an unstressed implies that there are two types of strains, namely
body, where the rst point, P1 is located at (x, y, z) the volume strain and the shape strain
and the second point (P2) at (x + dx, y + dy, (Fig. 2.4).
z + dz). Now, we let this body to deform as a
result of a stress system created within it. If the two
points (P1 and P2) were displaced from their 2.2.3 Common Types of Strain
original positions by equal displacements (D,
say), then it is considered that there is no strain Mathematical analyses of strain show that the
taking place. Strain occurs only when there is total strain of a three dimensional body, depends
variation of displacement of any point, within that on only six different derivatives of displace
medium, with respect to the others. In the lan ments. These strain components (eab), can be
guage of mathematics, we say that strain depends written down as follows (Richter 1958, p. 236):
on the derivatives of the displacement
components with respect to the chosen coordi exx = Dx / x
nates (x, y, z). The concept is claried in Fig. 2.3. eyy = Dy / y
ezz = Dz / z
exy = ( Dx / y + Dy / x) / 2
exz = ( Dx / z + Dz / x) / 2
P1 P2 eyz = ( Dy / z + Dz / y) / 2
(a)
Fig. 2.4 The two types of strain; volume and shape strains
2.2 Theory of Elasticity 27
Rigid bodytranslation and rotation represent (or compressional) stress. The longitudinal strain
cases of no strain, since no volume and no shape (e) is dened to be the change in length in a
deformation are involved. In Fig. 2.5, an ele certain dimension, of a body under stress relative
mentary cube (shown here in plan) is used to to its original length. For a rectangular lamina of
show simple types of elastic deformation (strain) dimensions (Dx by Dy), the longitudinal strains
and nostrain changes. (ex and ey) in the x and y directions are dened as
In general, an elastic body under stress can (Fig. 2.6):
experience two types of distortions; changes in The longitudinal strains can be extensional
volume and changes of shape. These changes, (tensile strain) or compressional (contraction
which occur as result of stress, are expressions of strain). Longitudinal strains (ex and ey) are
the physical properties of the stressed body. In its dened as:
simple form, elastic strain can be divided into
two main types. These are: the volumechanging ex Dx =Dx
strain (leading to body compression or dilatation) ey Dy =Dy
and the shapechanging strain (leading to body
shape distortion). where (Dx and Dy) are the changes in length in
x and y directions respectively.
The minus sign that appeared in the ey
2.2.4 The VolumeChanging Strain expression is entered to denote that the change
(Dy) is compression which is in opposite direc
The familiar example on this type of strain is the tion to the dilatation change (Dx), in the x
longitudinal strain of a body under an extensional direction.
Dy
y y
x x Dx
ex = Dx / x ey =  Dy / y
Fig. 2.6 Denition of the longitudinal strain as applied for a rectangular lamina of dimensions (Dx by Dy)
28 2 Seismic Waves
x x
2.2.5 The ShapeChanging Strain displacements. When the displacements are equal
a body may experience pure translation (rigid
As longitudinal strain gives expression for the bodytranslation) or pure rotation (rigid
volume changes resulting from stress application, bodyrotation), as shown in Fig. 2.8.
the shear strain gives the corresponding measure Rigidbody changes which do not involve
for the shape deformation. Using the example volume or shape changes, such as these, are not
above (Dx by Dy rectangular lamina). The shear considered to be elastic strains.
strain (also called angular strain) is considered to
be the average of the two angles by which two
neighboring sides rotate as a result of the shearing 2.2.6 The Cubical Dilatation
stress. Thus, the shear strain (exy) is dened as:
exy / b=2 A parameter, closely related to longitudinal strain
and of special importance in the theory of elas
where (/ and b) represent the angles of rotation ticity is the Cubical Dilatation (h). At a certain
of the two sides (Dx and Dy) brought about by point within a strained medium, this is dened as
the shear stress (Fig. 2.7). the fractional change in a unit volume surrounding
Since these two angles are very small (usually that point. Thus, for a three dimensional body with
so, in seismiceld conditions), they can be rep longitudinal strains (exx, eyy, ezz), the cubical
resented by their corresponding tangents, giving dilatation can be computed as follows:
where the angles (a & b) are in radians. For small strains exx, eyy, ezz (which is the
It should be emphasized that strain occurs case in seismiceld conditions), the products of
only if the body particles experience unequal these terms may be neglected giving the result:
2.2 Theory of Elasticity 29
h exx eyy ezz Dx =Dx Dy =Dy Dz =Dz For an isotropic body (physical properties are
independent of direction) and for an elastic body,
under small strain, strain varies linearly with the
The sign convention of (h) is negative for applied stress. This linear stressstrain relation
compression and positive for expansion strains. ship is governed by a wellknown mathematical
equation. It is the Hookes law.
strain
30 2 Seismic Waves
The linearity property governed by Hookes law 2.2.9 The Elastic Moduli
means that there is a proportionalityconstant for the
linear stressstrain relation for any particular body The elastic modulus of a body is the propor
under stress. Mathematical studies showed that, for tionality constant of the stressstrain linear rela
an isotropic body, two elastic coefcients are suf tionship. It expresses an important physical
cient (Richter 1958, p. 238). These are the Lames property which is the extent of resistance of that
coefcients (k & l), which are sufcient in char body to the applied stresses. Moduli of important
acterizing the elastic properties of a medium. practical applications are Youngs Modulus, bulk
By use of Lames coefcients (k & l), modulus, and shear modulus. These are dened
Hookes law can be presented in the following in the following discussions.
compact form:
(i) Youngs Modulus and Poissons Ratio
Let a simple tensile stress (Tx) be applied to
Tij = ij+ 2 eij an isotropic bar placed along the xaxis. This will
cause the bar to experience a longitudinal
extension (ex) in the xdirection and, at the same
where the symbols (i & j) take the values x, y,
time, it experiences lateral contractions along y
and z, and the term dij = 1 when (i = j), and
and zdirections. Being an isotropic body, the
dij = 0 when (i 6 j). T and e are the stress and
contractions in the y and zdirections (ey, & ez)
strain respectively.
are equal. These changes (expressed by the
This compact form of the Hookes law can be
strains ex, ey, & ez) are governed by the elastic
presented in the following explicit equations:
coefcients of the stressed body. The coefcients
which govern the stressstrain relation, in the
Txx = + 2 exx presence of the tensile stress (Tx), are Youngs
Tyy = + 2 eyy modulus (Y) and Poissons ratio (r).
Tzz = + 2 ezz For a onedimensional stress acting on a body
obeying Hookes law, Youngs modulus (Y) is
For pure shear strain (that is with no change in the proportionality constant in the linear relation
volume, for h = 0), the Law expresses the rela that connects stress (Tx) with strain (ex). The
tions for purely shearing strain, that is: relationship is:
V
A F
P
P x
VV
h
P
Fig. 2.11 An elastic cube under hydrostatic compression Fig. 2.12 An elastic rectangular block under shearing
forces force (F) acting on area (DA)
32 2 Seismic Waves
k rY=1 2r 1 r
2.3 Wave Motion Equation
l Y=21 r
If two neighboring points in a stressed medium
experience the same stress, no motion of one of
Other relations are presented in Table 2.3.
them will occur with respect to the other. How
ever, relative motion will take place when there
is a stress difference. In other words, motion
2.2.10 The Elastic Moduli
occurs when there is a stress gradient. This
Interrelationships
reminds us of an analogous case we met in the
creation of strain (see Sect. 2.2). The two cases
For a homogeneous and isotropic medium under
may be expressed as follows: Displacement
stress, the stressstrain relationship is linear
gradient is required to create strain and stress
within the elastic (Hookean) state. The propor
gradient is required to cause motion.
tionality constants are the elastic moduli or
2.3 Wave Motion Equation 33
2.3.1 OneDimensional Scalar Wave This strain (exx = Dx/x) is produced by the
Equation corresponding stress gradient (Txx/x).
By making use of the fact that the net force
In this section, we shall deal with the wave acting on any face is given by the stress acting on
motion equation which expresses the motion of a that face times the face area, we get the resultant
disturbance in one dimension. The disturbance in force (Fx) in the xdirection due to the stress
this particular case is the scalar quantity, the change (Txx/x) dx that occurred across the
cubical dilatation (h). distance (dx). This is computed as follows:
Let us consider an elementary parallelepiped
(of dimensions: dx, dy, Dz) located inside an Fx @Txx =@x dx dy dz
elastic isotropic medium (Fig. 2.13).
At each face of this elementary body, when it By applying Newtons second law of motion
is under elastic stress, there exist three stress we can express (Fx) in terms of mass of the
components: one is normal and two are tangen parallelepiped (dx dy dz times density q)
tial to the particular face. The three multiplied by acceleration (2Dx/t2) in the x
stresscomponents (Txx, Tyx, and Tzx) are acting direction giving:
on the face perpendicular to the xaxis. Under
dx dy dz q @ 2 Dx =@t2 @Txx =@x dx dy dz
elastic stressstrain conditions, each of these
components will have a gradient in the
xdirection (Txx/x, Txy/x, and Txz/x). For or:
a complete three dimensional state, additional
similar gradients occur in the other two directions q @ 2 Dx =@t2 @Txx =@x
(ydirection and zdirection).
In order to simplify the mathematical deriva This is the onedimensional (dimension, x in
tion of the equation of seismic wave motion, let a this example) wave motion equation which
plane compressional seismic wave to be describes particle motion (displacement, Dx) in
advancing in the xdirection. In this case the terms of the applied stress (Txx). However, the
three stresscomponents are reduced to only one motion can be expressed in terms of displace
component (Txx) creating the corresponding ment only. This is done by using the stressstrain
strain (exx). When a seismic plane wave propa linear relationship expressed by Hookes law
gates in the xdirection, the two faces (perpen equation for isotropic media (Txx = k h + 2l
dicular to the xaxis) of the parallelepiped will be exx). Substituting for (Txx), the previously
unequally displaced, and hence, it is subjected to derived wave equation becomes:
an elastic strain (exx) which is, by denition,
given by the displacement gradient (Dx/x). q @ 2 Dx =@t2 k@h=@x 2l@exx =@x
x
plane wave fronts
36 2 Seismic Waves
2.3.4 The P and SWaves 330 m/s in air, 1450 m/s in water, and (2000
6000) m/s in rocks.
As it is presented above, there are two types of A solid medium having its Poissons ratio
disturbance that can move in accordance with the equal to (1/4) is called Poissons solid (Sheriff
standard wave motion equation. These are the 2002, p. 266).
scalar cubic dilatation (h) and the vector shear
strain (w).
From the wave equation it can be shown that 2.4 Classification of Common
the disturbance (h) moves faster than the other Elastic Waves
disturbance (w). Thus, when the two distur
bances are generated by a certain source, the From analyses of stress and strain, we have seen
(hwave) arrives earlier than the (wwave). For that strain is, in general, made up of two types of
this reason the two waves are called Primary elastic disturbance; the cubic dilatation and the
(Pwave) and Secondary (Swave) respectively. shear strain. Solution of the equation of motion
It is to be noted that the ratio of the Pwave showed that each of these types of deformation
velocity (vp = [(k + 2l)/q]1/2) to the Swave travels through the medium with its own veloc
velocity (vs = [l/q]1/2) is equal to [(k + 2l)/ ity. The rst type of disturbance represents the
l]1/2. Using the relationship connecting (k) and moving volume strain and the second type
(l) in which k/l = 2r/(1 2 r), we can write: involves the shape strain. The rst type is
called Longitudinal, Compressional, or Primary
vp =vs 2 2r=1 2r1=2 wave (or just Pwave) which travels faster than
the second type which is called Transverse,
This formula clearly shows that the ratio of Shear, or Secondary wave (or just Swave).
the Pwave velocity (vp) to the Swave velocity These two types of waves (P and Swaves)
(vs) is function of Poissons ratio (r) only. belong to a class of waves (called body waves)
According to (Dobrin 1960, p. 18), Poissons because they can propagate through the interior
ratio (r) generally ranges from 0.05 to 0.40, of the earth body. This group of waves is called
averaging about 0.25 for hard rocks. With this so to differentiate them from another class of
value (r = 1/4), the velocity ratio (vp/vs) waves which move on and near the free surface
becomes 31/2 (=1.732). This means that Pwave of the medium, called (surface waves) which
moves with velocity which is about 1.7 times as include Rayleigh and Lovewaves. Classica
fast as the Swave moving in the same medium. tion of the common elastic (seismic) waves is
It is useful to note that Pwave velocity is shown in Fig. 2.15.
2.4.1 Body Waves Pwave is the fastest wave for a given medium
and, therefore, its arrival at a certain observation
Body waves are waves that can travel through an point is the earliest among the seismic
elastic materialistic medium in any direction. As wavetypes. This is a common observation of
they move, the waves may experience changes in seismologists working on analysis of earthquake
their energy level and in their travelpath geom seismograms. Propagation velocity (vp) of
etry subject to the physical properties of the Pwave depends on the medium density (q) and
medium. There are two subtypes of these waves; elastic properties (k & l) and it is given by the
the longitudinal and the transverse waves expression vp = [(k + 2l)/q]1/2.
(Fig. 2.16).
(ii) Transverse Waves
(i) Longitudinal Waves The travelling disturbance in this case is the shear
This type of waves is also known as compressional, strain or shape deformation. The medium
Primary, or just Pwave. The travelling disturbance which is traversed by this type of waves experi
in this case is volume deformation expressed by ences no volume changes. A consequence of the
the cubical dilatation (h) as dened above. shear strain (rotation of part of the medium) is the
The particles of the medium, traversed by a transverse displacement of the path particles
plane Pwave, vibrate about their neutral posi relative to the propagation direction. They are
tions in the direction of the wave propagation. also called (shear waves) or (Secondary, or just
The travel path consists of a sequence of alter Swaves).
nating zones of compressions and rarefactions A horizontally moving Swave, which is so
(Fig. 2.16a). This is the type of waves which is polarized that the particle motion is conned to
commonly employed in seismic reflection and vertical plane, is known as SVwave (Fig. 2.16
refraction exploration work. b). When the polarization plane is horizontal, it is
(b)
SH
propagation
direction
(c)
propagation
direction
38 2 Seismic Waves
called SHwave (Fig. 2.16c). The velocity of The main subtypes of surface waves are
Swaves, vs is given by vs = [l/q]1/2. In Rayleigh waves and Love waves (Fig. 2.17).
liquidmedia, where (l = 0), Swaves do not
(i) Rayleigh Waves
propagate.
Rayleigh waves, which were discovered by an
English scientist, Lord Rayleigh in 1885, are
usually developing at the free surface of a
2.4.2 Surface Waves
semiinnite solid medium. Its wave amplitude
decays rapidly with increasing depth. The trav
As it is implied by its name, surface waves are
elling disturbance in this case is a sort of com
waves that move on the free surface of the earth.
bination of particlemotions of both P and
The main features common among all surface
SVwaves. The particle motion, which has a
waves, observed on earthquake seismograms, are
retrograde elliptical orbit, takes place in a vertical
their relatively large amplitudes (high energy
plane parallel to propagation direction
content) and low frequencies when compared
(Fig. 2.17a). The minor axis of the elliptical orbit
with the body waves. In addition to that, they
is parallel to wave motion direction and it is
move with velocity which is generally slower
equal to twothirds of its major axis. Rayleigh
than body waves moving in the same medium. It
waves travel on the surface of a solid medium
is a common observation that the dispersion
with velocity of 0.92 of the velocity of Swaves
phenomena are more prominent in surface waves
moving in that medium (Bullen 1965, p. 90). In a
due to dependence of the velocity on the fre
sense, Rayleigh waves are similar to the familiar
quency of individual harmonic component.
water waves, with a fundamental difference, and
(a)
propagation
direction
(b)
surface layer
propagation
direction
Fig. 2.17 Particle displacementmode of a medium traversed by plane surfacewaves, (k) is wavelength. a Rayleigh
Wave, b Love Wave
2.4 Classification of Common Elastic Waves 39
that is the particle motion in case of Rayleigh Since they possess no vertical component,
waves describe an elliptical path whereas the Love waves are not detected by the geophone or
particlemotion path in case of water waves are by any suchlike verticalcomponent sensing
circular in shape. instrument.
In the case where the semiinnite medium is
overlain by a lowvelocity surface layer, Rayleigh
waves exhibit a phenomenon known as (disper 2.4.3 Seismic Noise
sion). Harmonic components of longer periods
(lower frequencies) travel faster. Consequently, Broadly speaking, the term (noise) used in seis
the Rayleigh wave seismograms would, in gen mology, is applied to all types of disturbance
eral, show decrease in period along the wavetrain. which may interfere with (and impose masking
Components of toolong wavelengths (too long effects to) the seismic signal of interest. In this
compared with the thickness of the surface layer) way, the concept of seismic noise bears a relative
penetrate deeper and travel with velocity of about implication. Thus, when the interest is focused
0.9 times the Swave velocity in the subsurface on reflected body waves, surface waves and other
material. The short wavelengths travel mainly in nonreflection waves (as direct and refraction
the surface layer with velocity of about 0.9 times arrivals) are considered to be the unwanted
the Swave velocity in the surface layer. troublesome noise. If the interest is in the
Surface waves, normally seen on shot records, refraction arrivals, reflection arrivals become
obtained in seismic reflection surveys, are com the unwanted noise. In the strict sense, however,
monly called (ground roll) and these are identied the ambient seismic disturbances (usually of
to be of Rayleighwave type. Sometimes, these are random energy distribution which form the
called pseudoRayleigh waves (Sheriff 2002). background of a distinct travelling signal) are
Groundroll waves are considered to be unwel considered to be the seismic noise.
comed noise and efforts are usually made to get rid Seismic noise has destructive effects on the
of them or at least minimize their masking effect seismic signals of interest. A signal recorded
caused to the seismic reflection signal. amid a background of noise is distorted and
weakened because of the interfering noise. Signal
(ii) Love Waves
resolution is badly affected with noise develop
This is the second subtype of surface waves which
ment. A measure for the signal resolution, called
was discovered, in 1911, by another English geo
the signaltonoise ratio (S/N ratio) is usually
physicist named A.E.H. Love (18631940). It
applied. It is dened to be the ratio between
develops only in cases where a solid elastic
signal amplitude detectable amid a background
semiinnite medium is overlain by a horizontal
seismic noise.
lowvelocity layer. Like SHwave vibration mode,
In exploration seismology seismic noise is
the particle movement is transverse and is conned
divided into two main types; coherent and inco
to the horizontal plane (Fig. 2.17b). Love waves
herent noise (Fig. 2.18).
travel by multiple reflections between the top and
bottom boundaryplanes of the surface layer. The (i) Coherent Noise
propagation velocity approaches Swave velocity Coherent noise is a seismic event characterized by a
in the subsurface medium for very long wave distinct apparent velocity and welldened onset.
lengths and to that of the surface layer for short In reflection seismology, coherent noise which
wavelengths (Dobrin 1960, p. 23). Love waves appear on shot records, are sourcegenerated seis
always exhibit dispersion. As in the case of Ray mic events. They are made up mainly of surface
leigh waves, Love waves propagationvelocity waves (ground roll) and airwaves which are of
increases with the period of the harmonic compo fairly narrow bandwidth with low frequency range.
nent. Again, the vibration amplitude decays expo Frequency content of this type of noise is typically
nentially with depth in the lower medium. below 20 Hz (Fig. 2.18).
40 2 Seismic Waves
amplitude
incoherent noise, random
noise (c), in relation to that
of the reflection signal (b)
0 20 40 60 80 100
frequency, Hz
(ii) Incoherent Noise at, in seismic data acquisition. Several ways and
Unlike coherent noise, the incoherent noise means are followed in the eldacquisition stage
consists of seismic events with unpredictable or in the following dataprocessing stage to get
amplitude and onset. This type of noise, which is enhanced S/N ratio. Suitable measures are
basically of random nature, forms the applied to the parameters of the seismic source
seismicenergy background of any seismic and detectors as well as those measures applied
shotrecord. In earthquake seismology it is in processing work, in order to attenuate these
commonly known as (microseisms), and in noises and enhance the S/N ratio.
prospecting seismology it is called (incoherent
background noise), or (ambient noise) as it is
sometimes referred to. In addition to the ran 2.5 Propagation of Seismic Waves
domness nature, the incoherent noise is charac
terized by a broad amplitude spectrum that Seismic waves are generated from a sudden
covers a wide range of frequencies compared change in the internal strain occurring inside an
with the nearly limited bandwidth of reflection elastic medium. The generating source may be
signals or coherent noises (Fig. 2.18). In the natural as in the case of earthquakes or articial,
geophysical literature we sometimes meet terms like exploding a charge of dynamite, as nor
like (white noise) indicating wideband noise, mally done in seismic exploration. All parame
and (red noise) for lowfrequency random noise. ters of an advancing seismic wave (waveform,
Intensive research work has been undertaken, speed, and travelpath geometry) may change
directed towards a greater understanding of the during the wave propagation. Form and mag
source and characteristics of the incoherent nitude of these changes depend on the physical
noise. It is now generally accepted that it is properties of the host medium. Whether the
generated as a result of external energy sources source is natural or articial, a seismic eld is
like wind movements, seawaves collisions with created when a sudden pressure pulse is initi
sea coasts, in addition to other various natural ated. The generated seismic energy moves away
and articial manmade activities. from the source zone in a form of a wave
Because of seismic noise which are unavoid motion propagation. Under these conditions
able seismic events which get recorded alongside (seismic energy source within an elastic med
the objective signal, the signaltonoiseratio ium), the seismic wave spreads out from the
(S/N) becomes an important parameter in signal source zone in every possible direction. A travel
detection studies. The S/N ratio is used as mea raypath, in a particular medium, is dened once
sure for the signal qualitylevel. Signal clarity the locations of both of the sourcepoint and
(S/N enhancement) is a central objective, aimed detectorpoint are dened.
2.5 Propagation of Seismic Waves 41
ray
wavefront
42 2 Seismic Waves
the wave front is a line drawn normal to the 2.5.4 The Concept of the Interface
spherical wavefront (circles on the surface
plane) at that point. Concepts of the wavefront The Interface is that boundarysurface separating
and rays are shown in (Fig. 2.20) for a case of two different media. As far as the changes (chan
dropping a pebble into a still pond. ges in spectral structure and propagation direction)
of seismic waves are concerned, two media are
considered to be different if both of the wave
2.5.3 Huygens Principle propagation velocity and the medium bulk density
are different. Since velocity is function of elastic
Huygens Principle states that each point on a coefcients, it can be said that density and elastic
wavefront acts as a source of a new wave which, properties are the factors which control the specic
in a homogeneous medium, generates a sec characters of the media. The parameter which
ondary spherical wavefront, the envelope of expresses the combined effect of velocity and
which denes the position of a wave generated at density is called (acoustic impedance) which is
some later time. dened to be the product of velocity by the density.
Huygens model of wave propagation requires To clarify the concept of the interface and the
that the secondary wavefronts are active only at roll of the acoustic impedance waves hitting an
the points where the envelope touches their sur interface let us consider a twolayer model which
faces. The wave energy is spreading out from the consists of two adjacent media (M1 & M2) of
primary sourcepoints in all directions, but their velocities and densities (v1 & q1) for medium (M1)
mutual interactions make the resultant distur & (v2 & q2) for medium (M2). The acoustic
bance zero everywhere except at the points impedances (z1 & z2) for the two layers are
where they touch the common envelope. (z1 = q1v1) and (z2 = q2v2) as shown in Fig. 2.22.
Applying the principle on planewave propaga In analogy to the role of electrical impedance
tion in a homogeneous, and in an inhomoge in the flow of electrical current, the acoustic
neous medium, is shown in Fig. 2.21. impedance expresses the extent of resistance the
Z1 = 1 V1 M1
Interface plane
Z2 = 2 V2 M2
Fig. 2.22 Concept of the Interface and denition of the Acoustic Impedance, z (=qv)
seismic energy meets when traversing a medium. place on raypath direction or on the waveform of
The higher the acoustic impedance, the lower the the travelling seismic pulse. In nature, however,
particle vibrationvelocity will be, and vice versa. the medium is far from this idealized form. In the
Acoustic impedance is measured by (kg s1 m2) solid crust of the Earth, it is commonly made up of
or by the equivalent (Ns m3) units. rock layers of varying physical properties and
At an interface, an incident seismic wave varying geometrical forms and sizes.
(normally a Pwave in seismic exploration work) In such inhomogeneous environments a
would, under certain geometrical conditions, give moving seismic wave would suffer from a num
rise to wave conversion in addition to reflection, ber of changes whenever it meets an interface
refraction, and diffraction. These cases shall be across which there is change in the properties of
dealt with in some details in the following the medium. In particular, changes in energy
discussions. content, waveform (spectral structure), propaga
tion velocity, direction of motion, and new wave
generation. These changes, are generated at the
2.5.5 Changes of Propagation interface planes dening the layer bounding
Direction at Interfaces surfaces (Fig. 2.23).
The common changes in raypath direction,
In an idealized homogenous and elastic medium, a which are of signicance to exploration seis
seismic wave propagates with no changes taking mology, are: reflection, refracted transmission
Layer1
Layer2
Layer3
Layer4
44 2 Seismic Waves
V1 V1 V1
V2 V2 V2
(refraction), and diffraction. These shapes of the of different density and elastic properties (dif
moving wave raypath occur at the boundaries of ferent acoustic impedances), four new wave
media having different seismic propagation phases are generated; reflected and refracted P
velocities (Fig. 2.24). and SVwaves. If, however the incident is
SHwave, the generated waves are only reflected
and refracted SHwave. The SVwaves, gener
2.5.6 Wave Conversion at Interfaces ated from an incident Pwave, (or Pwaves gen
erated from an incident SVwave) are called
When a seismic wave impinges on an interface (converted waves) (Fig. 2.25).
separating two media, which differ in acoustic An incident seismic wave onto an interface
impedances, the incident seismic energy is partly will be partly reflected and partly transmitted
reflected and partly transmitted with certain across the interface. In general, the interface will
waveform changes. When the raypath of an bring about wave conversion, reflection, trans
incident seismic wave is oblique, that is inclined mission, and diffraction. It should be noted here
with respect to an interface, new waves are that refraction is a special case of transmission.
generated. If, for example, the incident wave is Refraction (raypath bending) occurs only in the
Pwave (or SVwave) separating two solid media case of inclined incidence.
P SV SV SH
P SH
SV
P
M 1 , V1 , 1 M 1 , V1 , 1 M 1 , V1 , 1
M 2 , V2 , 2 M 2 , V2 , 2 M 2 , V2 , 2
P P
SV SV SH
V2 > V 1
Fig. 2.25 Wave conversion at an interface for three types of incident waves (P, SV, and SH waves). The symbol (h)
denotes angle of incidence
2.5 Propagation of Seismic Waves 45
2.5.7 Energy Partitioning incident waves (P, SV, SH) and impedances
and Zoeppritz elements (velocity and density) for each of these
Equations wavetypes. A typical set of Zoeppritz curves for
the case of an obliquely incident Pwave is
The mathematical expressions describing the shown in Fig. 2.26.
energy partitioning of an obliquely incident wave Referring to Fig. 2.26, it can be seen that, for
among the created converted waves, were normal incidence (angle of incidence = 0), no
derived rst by Knot (1899) and later on by Swave is generated and thus all the energy is
Zoeppritz (1907), but not published until (1919). shared by the reflected and transmitted (re
Using an approach (based on displacement fracted) Pwave. At a small angle of incidence,
computations), Zoeppritz has derived the equa the converted Swaves are of small energy level.
tions (commonly known as Zoeppritz equations) As this angle increases the generated Swaves
which express the relative energy partitioning as grow stronger at the expense of reflected and
function of angle of incidence and acoustic refracted Pwaves. At the critical angle of the
impedances of the media separated by the incident Pwave, the transmitted Pwave energy
involved interface. falls to zero, and at the same time, both of
In the geophysical literature, these equations reflected Pwave and reflected Swave grow
are presented in the form of curves for certain large. Buildup of energy of reflected Pwave, as
twolayer models with dened density and elas the critical angle is approached, is referred to as
ticity properties (see for example, Grant and (wideangle reflection). This phenomenon (in
West 1965, p. 54, Telford et al. 1990, p. 157, crease of reflection coefcient near the critical
Sheriff 2002, p. 401). A complete coverage of angle) is sometimes made use of in seismic
various types of incident waves, with different reflection exploration (Sheriff 1973, p. 241).
types of media, is found in (Ewing et al. 1957). Further, as the angle of incidence approaches
Because of the numerous possible grazing incidence (angle of incidence = 90),
parametervalues required to dene the behavior energy of the reflected Pwave increases and, at
of energydistribution as function of incidence grazing incidence (where there is no vertical
angle, many curves are required for the various component for the incident Pwave), the Swaves
cases. These cases represent selected types of disappear and no transmission process occurs
1.0
relative
energy transmitted Pwave
reflected Pwave
reflected SVwave
transmitted SVwave
Fig. 2.26 Typical Zoeppritz curves of energy partition Pwave at an interface separating two media of specied
ing as function of angle of incidence. The curves are for properties (sketched, based on Dobrin and Savit 1988,
the converted waves created by an oblique incident p. 43)
46 2 Seismic Waves
and consequently all the incident energy is con case of reflections from a given horizontal
ned to the reflected Pwave. reflector. There are, however, situations where
the angle of incidence does not vary with the
offset. Thus, in a multireflector case, the angle
2.5.8 Amplitude Variation of incidence (which is equal to angle of
with Angle of Incidence reflection for the same wavetype), varies with
reflector depth for a xed offset. Also, in certain
For nonnormal incidence, an incident Pwave cases, it is possible to get different offsets for a
leads to wave conversion in which both reflected xed value of reflection angle. These two cases
and transmitted P and Swaves are sobtained. which occur in multireflector situation are
The obliquelyincident wave energy is dis shown in Fig. 2.27.
tributed among all the converted waves in such a As expressed by Zoeppritz equations, the
way depending on the properties of the involved reflection coefcient shows variation with
media on both sides of the interface. According increasing angle of incidence (or with increasing
to Zoeppritz equations, the reflection coefcient offset). Depending on the distribution of the
is function of rock properties (density and elastic acoustic impedance on both sides of the inter
properties) in addition to the angle of incidence. face, the reflection coefcient can vary from
For a given reflector, the amplitude variation largenegative to largepositive values. This
with angle of incidence (AVA) of a reflected behavior (variation of reflection coefcient with
seismic wave is found to be dependent on Pois angle of incidence) can therefore be used as an
sons ratio as well as on impedance contrast indicator to predict lithological changes or type
across the reflection interface. In this way, the of fluid deposits.
parameter (AVA) possesses the same informa
tion contained in a combined P and Swaves
data. 2.6 Effect of the Medium on Wave
It is important to note that Zoeppritz equa Energy
tions give direct relation of amplitude variation
with angle of incidence (AVA) and not ampli Due to the earth ltering effect and other causes,
tude variation with offset (AVO). However, the wavelet generated by the source energy, is
offset is proportional to angle of incidence, in changed from its initial highenergy, impulsive
(a) surface
(b) surface
i1
i1
reflector1 reflector1
i2 i2
reflector2 reflector2
i3 i3
reflector3 reflector3
i1 > i2 > i3 i1 = i2 = i3
Fig. 2.27 Variation of reflection angle with reflector with increase of reflector depth for xed offset, and
depth for a xed offset and variation of receiver offset for b angle of incidence is constant for varying offset
a xed reflection angle. a Angle of incidence (i) decreases
2.6 Effect of the Medium on Wave Energy 47
Table 2.4 Factors 1. In the source zone 2. In the path zone 3. In the detector zone
contributing to energy
changes of a travelling EnergySource parameters Reflection coefcient Detector response
seismic signal Source coupling Geometrical spreading Detector coupling
Nearsource geology Inelastic attenuation Neardetector geology
Sourcegenerated noise Wave conversion Surface noises
Noises and interferences Noises and interferences Noises and interferences
reflector
medium. It is dimensionless quantity and inde parameter is dened to be the natural logarithm
pendent of frequency. The quality factor is inver of the ratio of two neighboring amplitudes of a
sely proportional to the attenuation factor (a). The gradually fading wavetrain. This is customarily
term 1/Q is called the specic dissipation. measured by the ratio of two amplitudes sepa
rated by one wavelength (Fig. 2.30).
By denition, the logarithmic decrement (d) is
2.6.3 Seismic Wave Energy given by:
Measurement
and the DB Unit d lna1 =a2 ln ear =ear k ln eak
function of the frequency component of the travelling wave, is moving. The wavetrain or the
travelling wave. Dependence of velocity on fre energy package (expressed by the envelope of
quency means that each frequency component of the wave train) is travelling with different
a seismic signal moves with its own velocity. velocity called (group velocity, U) as shown in
Thus a wave, composed of several Fig. 2.32.
frequencycomponents will experience The group velocity (U) is mathematically
componentseparation, and hence, related to the phase velocity (V) and wavelength
changeofform that occurs during travel. Dis (k) of the frequency component, by:
tortion of the waveform due to dependence of
the velocity on individual frequencycomponents U V kdV=dk
is called (wavedispersion).
The dispersion phenomenon leads to changing where, V, k, and dV/dk are average values for
of the shape of the wave train with travelled the range of frequencies making up the principal
distance. Each frequency component (that is, part of the pulse (Telford et al. 1990, p. 154).
each wavephase) moves with its own individual When the phase velocity (V) increases with
velocity (the phase velocity, V). This is the increase of the component period, it is termed as
velocity with which a given point, marked on the (normal dispersion), and in this case the group
2.6 Effect of the Medium on Wave Energy 51
time
0
velocity is less than the phase velocity (U < V). Dispersion phenomenon occurs in a disper
For the opposite case (inverse dispersion), it is sive medium, as when surface waves are travel
when phase velocity decreases with period we ling through a semiinnite medium which is
get (U > V). In the absence of dispersion, the overlain by a low velocity surface layer. Dis
two velocities are equal (U = V) and no distor persion of seismic body waves (P and Swaves)
tion to the wavform occurs. are too small to be detected in practice.
The Seismic Velocity
3
acoustic impedance for the most commonly coefcients increase in such a way that it offsets the
known rocktypes (AlSadi 1980, p. 70). effect introduced by the density increase. An
empirical relationship between Pwave velocity
(v) and bulk density (q) for the common sedimen
3.2.2 Elasticity and Density Effects tary rocks (sandstone, shale, limestone, anhydrite
) is given as follows (Gardner et al. 1974):
The wave motion equation of seismic waves
includes the velocity factor present in its mathe qv kv1=4
matical expressions. For a solid homogeneous
medium, the propagation velocity for P and where (k) is equal to 0.31 when (v) is in m/s and
Swaves (vp and vs) are functions of elastic constants equal to 0.23 when (v) is in ft/s.
of the medium (Lames elastic constants, k and l) as Density values for common rock types are
well as its bulk density (q). These functions are: found in Table 3.1 presented above.
vp k 2l=q1=2
3.2.3 Porosity and Saturation Fluid
and, Effects
1=v =vf 1 =vm involve the effects of the rock and fluid contents
individually. It has the form:
or, in terms of interval transit time, T (= 1/v), 1=v C1 C2
noted that only a small amount of gas, present in resulting in a velocity increase with both depth and
pores, produces sharp decrease in velocity, but geological age.
further increase in gas saturation produces only a Faust formula (Faust 1951), for a given geo
minor effect. This phenomenon was applied as a logical interval, can be presented in terms of the
tool for detection of hydrocarbon accumulation interval transit time, T (=1/v). This form is pre
in oil traps. Being more compressible than oil or sented as follows (Pennebaker 1968):
water, gas has the effect of lowering the velocity
much more than the presence of oil or water as Th Chn
interstitial fluids. Due to its effect on the elastic
properties of rocks, even a small quantity of gas or,
present in the pores, would result in large
log Th n log h log C
reduction in velocity (Sheriff and Geldart 1995,
p. 109). Clearly, this property has an important
where (C) is constant related to the rock lithol
application in hydrocarbon exploration.
ogy, pore pressure, and geological age.
For sandshale sequence, Pennebaker (1968)
found that the index (n) has a value of about
3.2.4 Depth and Geological Age
(1/4). This is considered to be representing a
Effects
geological section existing under normal
compactionpressure. As it is seen above, the
With depth of burial and with geological age, in
(Th) relation is linear when it is expressed in the
general, rock material becomes more compact.
logdomain. Thus, when a linear (log T log h)
This would result in increase in elastic constants
relation is obtained in a certain study, the geo
as well as in the density and, consequently,
logical section is considered to be under normal
increase in velocity. Dependence of velocity (v),
compaction pressure conditions. Deviation from
of a particular geological bed, on geological age
the linear relation (straightline relation) is nor
(g), existing at depth (h), is expressed by the
mally considered as indicative of abnormal
following empirical formula (Faust 1951):
pressure situation. This is another useful and
effective tool that can be used to investigate
vg; h kg hn
subsurface pressure conditions and as an indica
where (k) is constant (equal to 125.3 when depth tor for other stratigraphic features.
in feet, velocity is in ft/s, and geological age in
years). The index (n) is found in this study to be
equal to (1/6). In a later study (Gregory 1977) 3.2.5 Overburden Pressure Effect
made on sand and shaleformations data, the
index (n) is found to be equal to (1/4). A theoretical relationship developed by Gassman
Variation of velocity with the depth of burial (1951) showed that velocity of a rockbed model
was investigated by a number of geophysicists made up of spherical tightlypacked grains varies
(e.g. West 1950; Kaufman 1953; Acheson 1963), as (P1/6), where P is the applied compaction
conrming the general trend of velocity increase pressure. The similarity of this result to Fausts
with depth. empirical formula (Faust 1951) suggests that
As the function, v(h, g) is showing, geological increase in velocity (i.e. decrease in the interval
age bears a relationship with velocity similar to transit time) is due to increasing pressure
that of the depth factor. It shows that velocity is in imposed by the overburden rocks.
general increasing with depth and geological age. Compaction of a saturated porous layer is
The explanation is that, grain compaction increa function of the weight of the overlying material (i.e.
ses with increase of depth and geological age. In due to the overburden pressure, Po) and the pressure
consequence, elastic constants are increased of the saturation fluid (the pore pressure, Pp). Since
3.2 Factors Influencing Seismic Velocity 57
the effect of (Pp) in compressing the porous rocks is phenomenon. In case of anisotropic medium,
in opposite direction to that created by (Po), the velocity varies with changes in propagation
compaction net effect (compaction pressure, Pc) is direction. The main factors affecting velocity can
proportional to the pressure difference (Po Pp). be summarized as follows:
That is:
(i) Lithology
Pc Po Pp In general, igneous and metamorphic rocks
have larger velocity values than sedimen
As rock compaction process brings about a tary rocks. Subtypes of rock lithologies
corresponding change in the physical properties, it are characterized by wide ranges of
is expected that seismic velocity (or interval transit velocities which are overlapping with each
time) would vary with the applied compaction other. For example, the velocityrange for
pressure. Velocity increases with increasing sandstone (25) km/s is overlapping with
compaction pressure and decreases with increase that of shale (1.54.5) km/s, and so on.
of interstitial pore pressure. This velocitypressure (ii) Elasticity and Density
relationship forms the basis used in predicting Theoretically, velocity is directly propor
subsurface pressure conditions from seismic tional to elastic constants and inversely
velocity data. Use of seismic velocity as an indi with density. However, because of the
cator of pressure conditions was used by several greater effect of elasticity, it is observed
workers like Pennebaker (1968), Reynolds (1970), that velocity has an apparent direct pro
Louden et al. (1971), and Aud (1974). Departure portionality with density changes.
of the measured velocity of a given geological (iii) Porosity and Saturation Fluid
section from the normalcompaction trend serves Velocity is inversely proportional to
as indicator of abnormalpressure zone. porosity, and the change in velocity
depends on the type of pore fluid. Velocity
is lowered when the pores are lled with
3.2.6 Other VelocityAffecting water or oil, and more lowered when the
Factors pores are lled with gas.
(iv) Depth and Geological Age
In addition to the above mentioned factors, there In general, velocity increases as both of
are other less important factors which can cause the overburden depth and geological age
small effects on seismic velocity. Examples of increase. A powerlaw relationship is
such factors are pore shape, anisotropy, tempera connecting the velocity with the depth and
ture, and wavefrequency. It is found that velocity geological age. The power index for
decreases slightly with increase of temperature. It sandshale sequence is found to be around
is found that, for an increase of 100 C, velocity the value of (1/61/4).
decreases by about 5 % and that the velocity (v) Overburden Pressure
droprate is considerably more when the rocks are Due to compaction caused by the overbur
saturated with heavy crude oil or tar. Velocity in den weight, elastic constants as well as
water saturated rocks experiences sudden increase density get increased. There is a direct rela
as temperature is lowered to the freezing point tion between velocity and the net com
(Sheriff and Geldart 1995, pp. 120121). paction pressure which is equal to the
Velocity is practically unchanged over a broad difference between pore pressure and over
frequency range. Small change is observed in case burden pressure (geostatic pressure as it is
of bodywave dispersion (velocity variation with sometimes called). Velocity increases with
frequency). Lowfrequency components are fas increasing compaction pressure and decrea
ter because of frequencydependant absorption ses with increase of interstitial pore pressure.
58 3 The Seismic Velocity
vp vp vp vp
water saturated
oil
gas
oil saturated
Pore pressure fluid saturation temperature frequency
(vi) Other Minor Factors express the propagation velocity (V) as a math
Pore shape: Velocity decreases when ematical function of the traveldistance, V(x) or
pores are elongate. function of traveltime, V(t). For practical
Anisotropy: Velocity varies with changes applications, velocity is, more often, expressed as
in propagation direction. function of vertical reflection time, that is, as
Temperature: Velocity slightly decreases function of the twoway (reflection) vertical time
with increase of temperature. V(T). In a homogeneous, uniformly changing, or
Frequency: Velocity practically unchan layered medium, the distanceversustime func
ged with frequency. With normal disper tions can be linear, curved, or segmented
sion, lowfrequency components are respectively. The corresponding V(T) functions
faster because of frequencydependant can be plotted as curved or straight lines. Typical
absorption phenomenon. velocity functions are shown in Fig. 3.3.
Due to the geological complexity of the earth,
Main factors affecting velocity are shown in it is not, in practice, possible to express seismic
Fig. 3.2. velocityvariation in the form of an explicit
mathematical function. However, under certain
conditions, a geological medium is approximated
3.3 The Velocity Function by a dened set of specications that can allow
derivation of a mathematical function describing
Seismic propagation velocity, as we have seen the variation of velocity along the travelpath.
from the previous discussion, depends on a One of these models is one in which the instan
number of variables, like type of lithology, geo taneous velocity (V) increases linearly with depth
logical age, density, porosity, pressure, and depth. (h). This model is expressed by the linear func
In seismic reflection exploration work, velocity is tion; V(h) = V0 + kh, where (V0) is the velocity
usually presented as mathematical functions. The value at the surface (h = 0) and (k) is the
velocity function is taken to describe its variation velocitydepth gradient. The signicance of this
with reflection traveldistance (or, more often, relation is that it gives a close approximation to
with traveltime). the actual velocity variation observed in many
Since the real earth is not homogeneous sedimentarybasin areas (Dobrin 1960, p. 77).
(typically layered medium), seismic waves move Derivation of the raypath geometry (which turns
with different velocities in different parts of its out to be of circular form) and traveltime func
travelpath. In this case, it becomes very useful to tions, are given by several authors (see for
3.3 The Velocity Function 59
example, Nettleton 1940, pp. 257261 or Sheriff of the location in the travelpath of the moving
and Geldart 1995, pp. 9394). wave front. This means that at any point of the
travelpath, the velocity can be measured or
computed at that point. This is the (instantaneous
velocity) which can be dened to be the velocity
3.4 Types of Velocity Functions
of wave motion at a given point in a medium
traversed by an advancing seismic wave.
Unlike electromagnetic waves which are all of
Mathematically, the instantaneous velocity (v) at
one type (e.g. wave of visible light which is of
a point is given by the slope of the tangent to the
Swave type), seismic waves are of many types,
distanceversustime curve at that point (Fig. 3.4).
each of which has its own velocityvalue which
The instantaneous velocity, (v) of a wave
is dependent upon the density and elastic con
moving along a distance (x) is dened as:
stants, as mentioned above. Because of the
inhomogeneous nature of the crustal part of the v dx=dt
Earth (which is typically made up of layered rock
media), several types of velocity functions are
needed to express velocity variation as function
of travelled distance. Five velocitytypes are in
common use in seismic exploration. These are:
(i) Instantaneous Velocity (V) v = dx/dt
(ii) Interval Velocity (IV)
(iii) Average Velocity (AV)
(iv) Root Mean Square (RMS) Velocity (RV)
x(t)
(v) Stacking Velocity (SV)
The closest example to the instantaneous covering that distance. In seismic exploration
velocity is the velocity function in which the work, the average velocity is computed for a set
interval transit time (which is equal to the of layers, usually starting from the datum plane
reciprocal of velocity) is measured across each down to the required reflector level.
depthinterval of one meter down a drill hole. For a vertical travelpath, the average velocity
The resulting record (the sonic log) gives detailed of a geological section made up of (n) layers will
velocityinformation at points spaced by 1 m be given by dividing the total thickness hn (=
interval. This is a practical approximation of the Dh1 + Dh2 + Dh3 + + Dhn) by the traveltime
instantaneous velocityvariation with depth. tn (= Dt1 + Dt2 + Dt3 + + Dtn), that is:
AVn hn =tn
depth, h
3.4 Types of Velocity Functions 61
depth, h
of that layer respectively. Using the corre medium (medium made up of horizontal layers of
sponding travel times, (Tn) and (Tn1), the different interval velocities). A threelayer model is
required formula is: shown in the following Fig. 3.7.
3.4.4 Root Mean Square Velocity The root mean square velocity (RV) can be
calculated from the interval velocity data (IV),
Root mean square (RMS) velocity (RV) is dened down to the nth layer, by use of the following
as the square root of the average of the squares of the formula:
weighted intervalvelocities, where the weighting From the RMS velocity (RV), it is possible to
factors are the layer thicknesses or the interval transit derive the interval velocity (IV), given the two
times. In seismic exploration, the RMS velocity, like RMS velocities for the section, from the datum
other types of velocities, is usually computed for level down to the base (RVn) and down to the
vertical ray paths of waves traversing a multilayer top (RVn1) of that layer. Using the
Layer3
RV3 = [(IV21 t1+IV22 t2+IV23 t3)/(t1+t2+t3)]1/2
h3 , t3
depth, h
62 3 The Seismic Velocity
corresponding travel times, (Tn) and (Tn1), the gathertraces. This is one of the fundamental
required formula is (Dix 1955): processing steps (called NMO correction) which
For a given geological section, the RMS velocity bring about coherency of the reflection arrivals
is typically a few percents larger than the corre which, on stacking, give enhanced reflection
sponding average velocity (Sheriff 1973, p. 228). event. It is called stacking velocity because of its
role in enhancing the stacked reflection signal.
Since its direct role is in the NMO correction, the
3.4.5 Stacking Velocity term (NMO velocity) will be more appropriate
than the commonly applied term (stacking
Stacking velocity is the main velocity variable velocity). The CMP concept is explained in
that enters in the NMOcorrection formula (ex Sect. 4.4.
plained in Sect. 4.3). It is applied to remove the Role of the stacking velocity in enhancing the
timecontribution of the receivertosource dis reflection signal is schematically shown in
tance (called, offset) from the total reflection Fig. 3.9.
traveltime. This is claried in the following In dipping parallel reflectors (dip angle, h),
(Fig. 3.8). with parallel velocity layering, the stacking
Stacking velocity is determined by velocity velocity (calle it VS) is equal to RMS velocity
analysis technique whereby the time contribution (VRMS) divided by cosine of the dip angle (h)
of offset is removed before stacking of the CMP that is (Sheriff and Geldart 1995, p. 134):
T0 = 2h/V
Tx = (4h2 + x2 )1/2 / V
CDPgather stack traces other than the true propagation direction. In seismic
A B C exploration, the apparent velocity is often dealt
0 x
with in connection with the movement of plane
wavefront advancing along a path inclined with
respect to the ground surface. A plane wave
A approaching a ground surface with true velocity
(V), along a raypath making an angle (h) with the
B normal to the surface will have an apparent velocity
(Va) of its motion along the surface (Fig. 3.10).
During a time interval (t), the wave front
C moves a distance of (Vt) with the true velocity
(V), while, at the same time, the moved distance
T0 Tx at the surface (x) is covered by the apparent
velocity (Va), that is:
Fig. 3.9 Role of the stacking velocity in the NMO
correction of a CDP trace gather. In this model, three
stacking velocities were applied (cases: A, B, and C). The Va Dx=Dt
optimum velocity (case, B) gave the highest staked signal
hence,
VS VRMS =cos h Va V=sin h
This relation between the stacking velocity This relation shows that the apparent velocity
and RMS velocity can be applied by interpreters is always greater than the true velocity by a
in their interpretation work of seismic reflection factor depending on the angle of approach (h, in
data and in velocitychanges studies. this example). The apparent velocity approaches
Stacking velocity is sometimes called RMS innity when (h = 0), that is when the wave path
velocity, because stacking velocity is the nearest direction is perpendicular to the surface plane. In
in value to RMS velocity in a multilayer med any case the apparent velocity is always greater
ium. Like RMS velocity, stacking velocity is than the true velocity with which the wave is
slightly greater than the average velocity. approaching the horizontal surface plane.
Another point of interest which is related to
this subject is the apparent wavelength. Suppose
3.4.6 The Apparent Velocity that the measurements were taken for time of one
period (t = s), distance measured over the sur
The Apparent Velocity (Va) is dened as the face corresponding to time of one period will be
propagation velocity measured in the direction
x
Va t
Wave Vt
wave motion
front
direction
Fig. 3.10 Ray path geometry of a plane wave approaching the surface plane at an approach angle (h). Velocity in the
raypath direction (V), and velocity of wave front measured on the horizontal plane surface is the apparent velocity (Va)
64 3 The Seismic Velocity
one apparent wavelength (x = ka), hence (using 3.4.8 Representations of the Velocity
the relationship; k = Vs): Functions
depth
3.5.2 The UpHole Velocity Survey The electronic structure of the sonde is
designed in such a way that the output is made to
This method is applied to determine velocity be the difference in the traveltimes to the two
changes in relatively shallow depths, in the range receivers. The time difference, measured in
50100 m. It typical application is velocity timeunits per onefoot, called (interval transit
determination of the nearsurface weathered zone. time), is plotted (normally in microsecond units)
The same principles used in wellvelocity survey against depth to give the continuous wiggly
ing are applied in this method, but with the curve known as the (sonic log). Sonde basic
sourcedetector conguration reversed. In the structure is shown in (Fig. 3.14a).
uphole case, the sources (small dynamite charges) To avoid tilting and holeirregularities effects,
are placed inside the borehole at fewmeter spacing two seismicpulse sources are used instead of one,
and sequentially red. The detection system is making what is called a (boreholecompensated
placed on the surface at a location near the sonde) as shown in (Fig. 3.14b). Seismic pulses
wellhead. Velocity computation follows the same are emitted alternately from the two sources and
equations quoted above (Fig. 3.13). the transit times from the two oppositely traveling
refracted Pwaves are averaged electronically.
The borehole compensated sonde (BHC) gives
an average interval transit time which is plotted
3.5.3 Continuous Velocity Logging
on a paper strip. The produced log in this case is
normally referred to as BHC sonic log.
This method uses a recording system consisting
The BHClog is used in computing synthetic
of a boreholelogging tool, called the (sonde). In
seismograms, in identifying lithologies, and in
its standard form, it contains a seismicpulse
determining formation boundaries (Fig. 3.15).
source and two receivers, one foot apart, and
The interval transit time can be integrated down
source placed three feet from the nearest receiver.
the well to give the total travel time of the part of
Sound pulses, emitted from the source at uniform
the well for which continuous velocity logging
time intervals, are detected by the two receivers.
was conducted.
Due to the fact that Pwave velocity in the dril
ling fluid is lower than that of the rockmedium Seismic Data Analysis Methods
surrounding the well, the transmitted Pwave gets Reflection traveltime function, in a homoge
refracted at the wallside, moving into the rock neous and isotropic medium, is of a hyperbolic
medium and then recorded by the receivers.
sonic log
(a) (b) geological column transit time, sec
140 90 40
shale
S S
R limestone
R R
R
T0
T T
in the analysis is explained in more details in Since density variation is very small, com
Sect. 10.7.1. pared with velocity variation, that is putting
(q1 = q2), the direct (AV) relationship can be
3.5.7 Seismic Velocity Inversion obtained which is:
Table 3.2 Fields of application and precision assessment of various types of velocity (after AlChalabi 1979)
Velocity Main uses Precision requirements
Stacking velocity Stacking of seismic sections Modest to low
Migration processing Modest to low
Estimation of RMS velocity Dependant on situation
RMS velocity Migrationvelocity estimation Generally modest
Intervalvelocity estimation Dependant on situation
Averagevelocity estimation Dependant on situation
Interval velocity Lithologic and stratigraphic studies High to modest
Interpretation works Modest to low
Abnormal pressure detection High to modest
Ray tracing computations Dependant on situation
Migration processing Generally modest
Averagevelocity estimation Dependant on situation
Average velocity Depth conversion Generally modest
Interpretation works Modest to low
Precision requirements: high = 0.11.0 %, modest = 15 %, low >5 %
static and dynamic (NMO) corrections. Amplitude conversion and in mapping structural and strati
compensations (as in geometrical spreading and graphic features.
inelastic absorption) and seismic migration Summary of elds of application and preci
depend on velocity. In interpretation activities, sion assessment of the various velocitytypes are
velocity has a fundamental role in timetodepth given in the following Table 3.2.
Seismic Wave Reflection
and Diffraction 4
In the real Earth Crust, a seismic wave may meet A diffraction event bears a marked relation
a variety of geological changes. Typically, the with reflection events since both are types of
media traversed by seismic waves are made up of seismic energy generated from an intervening
layered rock formations of different physical reflector. A plane surface (surfacereflector)
properties and different geometrical shapes. In causes reflection and point obstacle (point
such environments, some of the seismic wave reflector) causes diffraction.
energy gets reflected from interfaces or diffracted
from structural obstacles. The rest of the incident
wave is transmitted through the interfaces with
their raypaths being bent (refracted) in case of 4.2 Wave Changes at Reflection
inclined incidence and with no bending when Interface
raypath is perpendicular to an interface.
In this chapter, discussion shall deal with the The reflection process involves two main types
reflection and diffraction of seismic waves since of changes. These are: change in the propagation
they are physically more closely related to each direction and change in the energy content
other. Transmission with its special case, the (Fig. 4.2).
refracted transmission (refraction) shall be dealt The two types of changes occurring in the
with separately in the following chapter. reflection phenomenon involve the wave energy
content and its travelpath geometry. These are:
Fig. 4.1 The four principal types of wave arrivals (direct, reflection, refraction, and diffraction) representing the
commonly recorded seismic events
SV
amplitude) compared with the incident wave. Sometimes, the reflection coefcient is dened
The ratio of the reflected amplitude to that of the in terms of energy ratios instead of amplitude
incident represents the reflection coefcient. ratios. With this approach, the coefcients are
Zoeppritz equations provide the mathematical expressed by squares of the amplitudes.
expressions for the reflection and transmission By applying the stress and strain continuity
coefcients of all of the produced waves (two conditions at the interface level (z = 0), it is
Pwaves and two SVwaves in this example). possible to derive expressions for the reflection
The important feature here is that the reflection and transmission coefcients in terms of the
and transmission coefcients are functions of contrast in acoustic impedance (see Sheriff and
both of the angle of incidence and the contrast in Geldartl 1995, p. 76). The reflection coefcient
acoustic impedance existing across the interface. (R = Ar/Ai) and transmission coefcient
The fundamental theoretical principles upon (T = At/Ai) for the case of normal incidence, are
which solution of Zoeppritz equations is based given in terms of the contrast in the acoustic
on, is the fulllment of the conditions which impedance (Z) of the two media on either side of
require that all normal and tangential stresses and the interface. Thus, by denition, the reflection
displacements at the interface are continuous. For and transmission coefcients (R and T) are:
displacement continuity, the sum of normal and
tangential displacementcomponents (at the R Ar =Ai q2 v2 q1 v1 =q2 v2 q1 v1
interface) in the rst medium must be equal to Z2 Z1 =Z2 Z1
the corresponding sum in the second medium. T 1 R At =Ai C=A 2q1 v1 =q2 v2 q1 v1
Concerning stress, it is required that, the sum of 2Z1 =Z2 Z1
normal and tangential stresscomponents are
similarly equal. Another important note in this
If (R) is positive, compression displacement is
contest is that Zoeppritz equations are based on
reflected as compression because, by reflection in
the assumption that all the involved waves are
this case, both of the wavemotion direction and
pure sine waves of same frequency which drops
particle displacement are reversed. If, however,
out of the equations (Richter 1958, p. 670). This
(R) is negative, a compression is reflected as
means that no frequency change taking place in
dilatation, meaning that we get (180) phase
reflection and transmission processes.
change in this case.
It is evident from these equations (case of case where the acoustic impedance in one of the
normal incidence) that the reflection coefcient two adjacent media is approaching zero or innity.
depends only on the contrast in the acoustic Thus, when (Z1) is very small compared with (Z2),
impedance (Z2 Z1). The greater the contrast, that is when (Z1) approaches zero, (R) approaches
the larger reflection coefcient will be. This unity (R = 1). In this case, the interface is consid
implies that no reflection occurs from an inter ered to be ideal reflector since this means that all of
face across which the acoustic impedance the incident energy is reflected and no part of it is
assumes the same value, even if either velocity or transmitted through. A good approximation of this
density varies individually across the interface. kind of situation is the earth freesurface which is an
The reflection coefcient ranges in value from interface between the upper airmedium and the
(1) to (+1) depending on the acoustic impe lower rockmedium. Because of the vast difference
dances (Z1 and Z2). When the acoustic impe between the acoustic impedances of these two
dance (Z2) in the second medium (medium in media, the reflection coefcient approaches the
which transmission occurs) is greater than that in value of (+1) for a source located in the air and (1)
the rst medium (medium of incident wave, Z1) a for a source located inside the rockmedium. The
compression displacement is reflected as com interface in both cases is equally efcient in
pression, and the reflection coefcient is positive. the reflecting process, but in the second case there
In the opposite case, that is when (Z1 > Z2), a occurs a polarity reversal, or phase change of (p) for
compression is reflected as rarefaction (phase an incident sine wave.
change = p) and the reflection coefcient
becomes negative. Reflection coefcient is zero
when the two acoustic impedances (Z1 and Z2) 4.2.3 Geometry of Reflection
are equal, which means that there is no interface from Dipping Reflectors
existing in the way of the incident wave.
Another useful feature which may be deduced The general case of reflection geometry is the
from the reflection coefcient expression, is the case of inclined (dipping) reflector (Fig. 4.5).
(b) x 0 x
Xm
traveltime
apex
curve
T(x)
4.2 Wave Changes at Reflection Interface 75
V1 horizontal reflector
V2
x 0 x
apex
traveltime
curve
T(x)
DT x2 =2T0 V2
4.3.1 The Normal MoveOut (NMO)
Concept This is considered as an accepted approxi
mation since (x/VT0 1), which is usually the
The Normal MoveOut (NMO) is dened as the case in seismic reflection exploration. In this
difference (T) between reflection traveltime mathematical process, the exact form of (T) is
(Tx) and the twoway vertical traveltime (T0). transformed from its hyperbolic function to the
For a single horizontal reflector found at the base approximate form which is a parabolic function.
of a homogenous layer (of constant velocity), the Both forms of these two equations show that
NMO parameter (T) is given by Fig. 4.8. (T) is function of three variables; the receiver
4.3 The NMO and DMO Concepts 77
surface
reflector2
reflector3
Tx
T
z
reflector
Fig. 4.8 The NMO concept. a Twoway vertical raypath. b Twoway slant raypath. c Reflection arrival times
corresponding to the raypaths of a and b. The NMO (DT) is shown as the difference, (Tx T0)
T T T
x T0 V
Fig. 4.9 NMO (DTfunction) direct variation with (x) and inverse variation with both T0 and V
offset (x), velocity (V), and the twoway vertical For a given offset (x), the time (Tx) can be
time (T0). The proportionality is direct with (x2) readily measured for a certain reflection event
and inverse with both (V) and (T0). These appearing on the seismic trace recorded at that
changes are shown as follows Fig. 4.9. offset. From these data, the NMO (T) can be
78 4 Seismic Wave Reflection and Diffraction
Tx Tx
R S R R S R
surface
reflection
raypath
S R surface S
S3 S2 S1 R R1 R2 R3
reflection raypath
horizontal reflector
CDP
horizontal reflector CRP
DP CMP
RP
MP Fig. 4.13 Denition of the common points (CDP),
(CRP), and (CMP), which coincide on each other in case
Fig. 4.12 Denition of the depth point (DP), located of horizontal reflector
vertically below the midpoint of the sourcereceiver
distance. It is the reflection point (RP), or midpoint (MP)
surveying. Application of the concept usually
results in great enhancement of the reflection
These points are described as common points signal by summing together (stacking) the
(CDP, CRP, CMP) in cases where multi CDPGather traces.
reflections occur for the particular point. In other
words, a point is described as common when it
becomes common to more than one reflection. 4.4.2 CDP, CRP, and CMP in Case
of Dipping Reflector
4.4.1 CDP, CRP, and CMP in Case For a horizontal plane reflector, the points (DP,
of Horizontal Reflector RP and MP) coincide at a point which is located
vertically below sourcetoreceiver midpoint. In
By use of certain sourcereceiver layout, it is the case of dipping reflector, however, the
possible to shoot a number of shots such that the reflection point gets shifted updip by a distance
reflection points resulting from this number of depending on the offset as well as on the dip
shots coincide on each other, that is one angle of the reflector. The updip shift is given
pointlocation will serve as common point to all by:
of the implemented shots. In this type of repeated
shootingspread, the reflection point becomes DL x2 cos h sin h=D
known as common depth point (CDP), common
reflection point (CRP), or common midpoint where (L) is the updip shift of the reflection
(CMP). These points are shown in Fig. 4.13. point (RP), (x) is half the sourcereceiver offset,
The four reflection raypaths, shown in (h) is reflector dipangle, and (D) is the length of
Fig. 4.13, have one common depth point (the reflection raypath which is normal to the
CDP), will each produce a seismic trace. These reflector (Deregowski 1986). The raypath
traces belong to the same depth point (CDP) and geometry and location of the reflection point
thus they form a group of traces called the (RP) are shown in Fig. 4.14.
(CDPGather). In this case (case of horizontal In case of repeated shooting (as it is done in
reflector) the CDP and the other points (CRP and the usual seismic reflection proling surveying),
CMP) will coincide on each other. the reflection points (RPs) get dispersed
Ever since the CDP concept was introduced along the dipping reflectorplane. For a given
by Mayne (1962, 1967) it has been applied on a reflector, the amount of updip shift (L)
routine basis in execution of seismic reflection increases with the square of the receiveroffset
4.4 The CDP, CRP, and CMP Concepts 81
RP3
RP2
RP RP1
82 4 Seismic Wave Reflection and Diffraction
S
S3 S2 S1 R R1 R2 R3
CMP
CRP
Fig. 4.16 Denition of the common reflection point (CRP) obtained from multireflection experiment. For dipping
reflector, the (CRP) is located at one reflection point which is coincident with the end point of the normal from the CMP
diffracted wave
spherical wave fronts
in every possible direction giving a phenomenon this type is normally referred to as a point
called (diffraction) and the wave which leaves the diffractor.
obstacle after incidence is called (diffracted The diffraction phenomenon can be explained
wave). by considering the diffraction point as a
A closely related term is (wave scattering) pointsource which, upon being hit by an incident
which is used to describe diffracted waveeld wave, becomes an activated source that radiates
caused by small structural irregularities, as for waves in all directions. According to Huygens
example found with seismic energy reflected principle, an obstacle hit by an incident wave
from rugged basement surfaces. becomes an energysource from which seismic
waves are generated and transmitted in all possible
directions. If a plane seismic wave, for instance hits
4.5.1 The PointDiffractor a pointdiffractor embedded in a homogeneous
medium, a diffracted wave with spherical wave
A structural obstacle, having radius of curva fronts will be generated. The diffracted wave will
ture which is shorter than that of the incident move away from that sourcepoint causing inter
wavelength, acts as a diffractiongenerating ference with the incident wavetrain and all other
point (or diffraction point). An obstacle of waves that may be coexisting at the time (Fig. 4.17).
4.5 The Seismic Wave Diffraction 83
reflected wave
incident wave
terminating reflector
terminating reflector
fl t point
pointdiffractor
Point
point diffractor
diffractor
diffracted
spherical
wave front
84 4 Seismic Wave Reflection and Diffraction
0 0
Tr Tr Td Td
depth
point source
x
T0
Tx
seismic image
diffraction
time hyperbola
edge, we have for the same offset (x), the travel and,
time function (TDx) for diffraction, and (TRx) for
the reflection, are given by: DTd x2 =T0 v2
h i1=2 giving,
TDx x=v2 T0 =22
h i1=2 DTd 2DTr
TRx x=v2 T20
This shows that the normal moveout of a
and, wave diffracted from a terminating reflector is
approximately equal to double that of a wave
h i1=2
DTd TDx T0 x=v2 T0 =22 T0 =2 reflected from the same interface at the same
h i1=2 offset. This important feature is used in dis
DTr TRx T0 x=v2 T20 T0 criminating diffraction from reflection events.
where, (Tr) and (Td) are the normal moveout 4.5.5 The Diffraction Hyperbola
for the reflection and diffraction respectively, and
(v) is the propagation velocity. For small The seismic image of the diffraction arrivals is (for
offsettodepth ratio, that is for (x/vT0 < 1), these constant velocity) a hyperbolic curve centered
equations can be approximated by: about the diffraction sourcepoint. A diffraction
event such as this is expected to appear on a seis
DTr x2 =2T0 v2 mic section made up of zerooffset traces. The
86 4 Seismic Wave Reflection and Diffraction
diffraction point
diffraction hyperbola
seismic stack section is effectively a section of of their interferences, these arrivals are distorting
zerooffset traces (Fig. 4.21). or masking the reflection events and causing
Diffraction hyperbolae are often observed on smearing effects at the faultzone leading to
seismic stack sections, indicating pointsource decrease in faultresolution. Example of the
diffractions, as those generated by faults and interferences introduced as a result of fault
pinchouts. A seismic diffraction event (diffrac generated diffraction hyperbolae are shown in
tion hyperbola) appearing on seismic stack sec Fig. 4.23.
tion can be considered as the seismic response of Although diffraction events (diffraction
a depth section of a homogeneous medium con hyperbolae) are introducing masking effects to
taining a pointsource diffractor (Fig. 4.22). the primarily targeted reflection events, they
(diffraction events) can help interpreters in
identifying faults and other diffractionsources.
4.5.6 Distortion Effects of Diffraction In general, diffraction events on stack sections
are considered to be unwanted distortive events
In nature interfaces are not always plane and which must be removed or at least attenuated as
continuous surfaces. There are cases where these much as possible. The common way to remove
surfaces are irregular as the surfaces of reef the diffraction events (diffraction hyperbolae) is
bodies or discontinuous as the faulted beds or by application of the seismic migration which is
pinchouts as found with angular unconformities. one of the principal steps applied in processing of
When a seismic wave is incident on such sub seismic reflection data.
surface features, diffracted waves are generated
and transmitted through the medium, interfering
with, and distorting other coexisting waves such 4.5.7 The ExplodingReflector Model
as reflection events, the main objective in normal
seismic exploration. In accordance to Huygens principle, the
On a seismic stack section, which is effec exploding reflector model is a model in which a
tively made up of zerooffset traces, the diffrac reflection interface is considered to be formed of
tion arrivals (diffraction event) appear as a innite number of closely packed diffraction
hyperbolic curve whose apex is coincident with points and that these points explode at a given
the causing diffraction point. Faulting is one of starttime, generating seismic waves. The
the principal sources of diffraction waves seen in exploding reflector modeling process requires
stack sections. A reflector termination caused by that each point of the refelector is sending seis
faulting, generates diffraction arrivals. Because mic rays, in every direction (exploding) at a
4.5 The Seismic Wave Diffraction 87
seismic
image
diffraction hyperbolae
common start time. The moving seismic ampli wave fronts of waves diffracted from closely
tude is considered to be of magnitude propor packed diffraction points (making up the reflec
tional to the normalincidence reflection tion surface), is called exploding reflector model.
coefcient. The resulting wave eld generated according
The generated waves are made to propagate to to the explodingreflector model shall propagate
the surface with velocity half that of the actual along the normalincidence raypaths. The con
velocity. With this velocityvalue the oneway cept is schematically shown in Fig. 4.24.
traveltime to the surface becomes equal to the Compared with the conventional record sec
twoway reflection traveltime for zerooffset tion, the section produced from the exploding
receivers placed on the surface (Sheriff 2002, reflector model, is the same as the conventional
p. 127). Consideration of the reflection process as section having propagation velocity equal to half
being formed from the interactions of seismic of the true velocity (Fig. 4.25).
88 4 Seismic Wave Reflection and Diffraction
Application of the exploding reflector concept in of migration (both of the pre and poststack
computing the seismic reflection section is a migration processing) using the explodingreflector
directmodeling (forwardmodeling) process. concept. This approach proved to result in great
Algorithms have been developed for computations reduction of the migration computational cost.
Seismic Wave Transmission
and Refraction 5
The term (transmission) is customarily used to layers down to a reflector and up to the surface.
indicate the general case of wave propagation Among other factors, reflection and transmission
where the moving seismic wave crosses an coefcients cause attenuation to the wave energy
interface whether the incidence is normal or as it is experienced by the recorded function r(t).
inclined. When an obliquely incident wave is It should be remembered that there is a reflection
transmitted across an interface, it is bent at the and transmission processes taking place at each
interface towards, or away from, the normal at interface existing in the traversed geological
the point of incidence. This is the well known section (Fig. 5.1).
phenomenon of (refraction). A refracted wave is,
therefore a transmitted wave resulting from an
inclined incidence. No refraction occurs when 5.1.1 Transmission Coefficient
the wave is incident perpendicularly at an inter at Normal Incidence
face. Thus, it can be sayed that the refraction
phenomenon is a special case of transmission. It The transmission coefcient (or transmittance,
(refraction) occurs only with normal incidence. as it is sometimes called) is dened to be the
ratio of the amplitude of the transmitted wave to
that of the incident wave. Sometimes this is
5.1 Seismic Wave Transmission expressed in terms of energy instead of ampli
tude. At normal incidence, the transmission
When a seismic wave hits an interface, part of the coefcient (T), expressed as a ratio of the
wave energy is reflected and the rest is transmitted transmitted amplitude (At) to the incident
across that interface. The amount of energy which amplitude (Ai), thus:
is reflected from an interface is determined by the
reflection coefcient and what is left from the T At =Ai
incident energy is transmitted across the interface
into the second medium. At the interface the As it is with the reflection coefcient, the
raypath is bent if the incidence is oblique and transmission coefcient is function of the
continues with unchanged direction when the acoustic impedances of the two media separated
incidence is normal to the interface. by the involved interface. For an interface sepa
In the real layered Earth crust, the initiated rating two media of acoustic impedances (Z1 and
seismic energy, the source function s(t), is Z2), the transmission coefcient (T), is related to
recorded as r(t) after it penetrates all the rock the impedances by:
R4 T12 1 R 1 R 1 R2
Fig. 5.1 Waveattenuation due to reflection and trans And in terms of impedances it is given by:
mission processes which are normally occurring in a real
layered rock medium. s(t) and r(t) are source receiver T12 4 Z1 Z2 =Z2 Z1 2
timefunctions respectively
The twoway transmission coefcient (T12) is
useful in computing the effective attenuation
T 2Z1 =Z2 Z1
factor of a wave reflected from a subsurface in
It is apparent from this formula that as the terface after being transmitted through several
contrast in the acoustic impedances is smaller the layers during its total reflection travel path.
transmission coefcient becomes greater, and in
the limit when they are equal (Z1 = Z2), trans 5.1.3 Attenuation Due to Reflection
mission coefcient (T) becomes unity and all of and Transmission
the seismic energy shall pass through and no
reflected energy will take place. For an nlayer medium, the wave which is
For the case of inclined incidence, the trans reflected from the nth interface and received at
mission coefcient, like the reflection coefcient, the surface would have crossed the (n 1)
depends on both of the angle of incidence and on interfaces twice. Assuming normal incidence, the
the contrast in the acoustic impedance of the effective attenuation factor (RTn) of a wave
adjacent media. Unlike reflection, however, there reflected from the base of the nth layer (nth
is no special situation whereby phase change subsurface interface) and transmitted through
occurs. (n 1) interfaces will be given by:
RTn An =Ai
5.1.2 The TwoWay Transmission
Coefficient By use of the concept of the twoway trans
mission coefcient, it is possible to derive a for
The transmission coefcient (T) is related to the mula that computes the net attenuation due to the
reflection coefcient (R) by the expression (see combined effects of the reflection and transmis
Sect. 4.2.2): sion coefcients. In a multilayer geological model
5.1 Seismic Wave Transmission 91
R3
R4
A1 Ai R1
5.1.4 Role of Transmission in Seismic
A2 Ai R2 1 R21 Exploration
A3 Ai R3 1 R21 1 R22
... Seismic wave transmission surveys make use of
seismic waves which are directly propagating
An Ai Rn 1 R21 1 R22 1 R23 . . . 1 R2n1 from sources to receivers. The normal direct
wave, which is always recognized on seismic
A general recursive formula can be derived by reflection and refractionshooting records, is a
dividing the amplitude of the nth reflector by that typical transmission method used to calculate the
of the (n 1) reflector, giving: velocity of the surface layer. In exploration
seismology, there are several transmissionbased
An An1 Rn Rn =Rn1 1 R2n1 ; n2 methods which are applied for specic explo
ration purposes. Uphole surveying, well velocity
Repeating, this formula is valid for (n 2), surveying, vertical seismic proling (VSP) fan
and for (n = 1), we have A1 = Ai R1. shooting, and seismic transmission tomography,
In conclusion the effective attenuation factor are typical examples of exploration practices
RTn (=An/Ai) due to reflectiontransmission employing the seismic wave transmission phe
process for a wave reflected from the nth nomenon. The uphole surveying and that of
reflector of a multilayer geological section, is wellvelocity are dened herebelow (Fig. 5.3).
92 5 Seismic Wave Transmission and Refraction
seismic
source detector
i r
M1 (V1 ) M1 (V1 ) M1 (V1 )
M2 (V2 ) M2 (V2 ) M2 (V2 )
rr i
i<r i>r
i=r=0
Fig. 5.4 Occurrence of refraction in case of oblique direction reversed. The velocities of the two media are V1
incidence. a Normal incidence, no refraction, b oblique and V2, where V1 < V2
incidence, from medium M1 to medium M2, c incidence
i i
medium1 velocity v1
medium2 velocity v2
r
refracted
wave
SV
P
P
rsv1
Fig. 5.7 An obliquely incident Pwave and the generated four wave phases: reflected and refracted P and SVwaves
5.2 Seismic Wave Refraction 95
an interface. If medium2, for instance, is liquid, incidence angle (i) increases, and when it reaches
no transmitted SVwave is formed. a value of (i = ic), refraction angle reaches the
Using the symbols shown in Fig. 5.7, the value (r = 90). The angle of incidence (ic) for
generalized Snells law is quoted as follows which the angle of refraction (r) becomes (90) is
(Sheriff 1973, p. 200):
(sin ip1)/vp1 = (sin rp1)/vp1 = (sin rsv1)/vsv1 = (sin tp2)/vp2 = (sin tsv2)/vsv2 = k
incident P reflected P&SV refracted P&SV
5.2.2 The Critical Refraction and In this state, where the incident angle is equal
Head Wave Generation to the critical angle (ic), a special refracted wave
(called the head wave) is generated. This criti
Consider a geological model made up of a med cally refracted wave travels along the interface
ium (M1) of velocity (v1) overlying a medium with the velocity (v2) of the second medium
(M2) of highervelocity (v2). An obliquely (M2). This wave is refracted back to the earths
incident wave in the rst medium shall refract surface at the same angle (ic) and with the
on hitting the separating interface. In such a propagation velocity of the upper medium (v1).
model, the refraction angle (r) increases as the Development of head waves is shown in Fig. 5.8.
90
96 5 Seismic Wave Transmission and Refraction
Head waves have great practical importance the direct wave. The sourcereceiver distance
as they form the fundamental basis of seismic (0 to xcr), at which the refraction wave overtakes
refraction exploration. the direct wave, in the same layering setup, is
called the (crossover distance).
The velocities of the two involved layers (v1 and
5.2.3 RayPath Geometry
v2) are computable from the slopes of the traveltime
and TravelTime Curves
lines (as seen from the gure above). The velocity is
the inverse of the slope of the timedistance line.
On seismic shooting records, three main wave
arrivals are usually observed. These are the direct,
refracted, and reflected waves. Their traveltime 5.2.4 Refraction TravelTime
functions are linear for the direct and refracted Function
arrivals, but hyperbolic for the reflected wave. At
the critical distance (xc), where the incidence is at A traveltime function of a moving wave
the critical angle (ic = sin1(v1/v2)), a head wave is expresses the mathematical relation between the
generated which is refracted to surface at an angle time covered by the moving wave and the
equal to the critical angle. The distance (0 to xc), no sourcereceiver distance. This is also depending
refraction arrivals exist, hence it is a (shadow zone) on the velocity values of the two involved media.
for refracted waves (Fig. 5.9). Two cases are presented here; these are cases of
For a given interface separating two constant horizontal and dipping plane interfaces.
velocity media, the refraction traveltime function
is a straight line tangent to the reflection hyper (i) Case of Horizontal Refractor
bola at the point where both of these waves arrive Consider rst a simple onerefracting horizontal
at the same time. The distance at which reflection interface separating two layers of velocities (V1
time equals the refraction time, which occurs at and V2) as shown in Fig. 5.10.
the critical angle, is called the critical distance. Referring to this gure, the refraction raypath
Due to its shorter travelpath, the direct wave, consists of three segments, AB, BC, and CD. The
at small receiver distance, arrives earlier than the refraction travel time T(x) is given by
refraction wave. However, after some time, the AB/V1 + BC/V2 + CD/V1. That is:
faster refraction wave catches up and surpasses
Tx 2z=V1 cos ic x2z tan ic =V2
reflection curve or, (using sin ic = V1/V2 from Snells law and the
Tx refraction curve sinecosine relationship)
Tx x sin ic 2z cos ic =V1
direct curve
or,
1=2
T0 Tx x=V2 2z V22 V21 =V1 V2
x
0 xc xcr where (ic) is the critical angle, and (z) is depth of
the interface. The two velocities, (V1 and V2) are
velocities of the two layers. These are equal to
z ic ic the reciprocals of the slopes of the direct and
M1, v1
refracted traveltime curves respectively.
head wave M2, v2 It is readily noted that refraction traveltime
function is linear in (x), and its curve is straight
Fig. 5.9 Raypaths and traveltime curves of direct,
criticallyrefracted, and reflected waves, with critical line of slope equal to (1/V2) and its intercept (T0)
distance (0 to xc) and crossover distance (0 to xcr) is given by:
5.2 Seismic Wave Refraction 97
B ic z2
C V1
V2
1=2
T0 2z V22 V21 =V1 V2 : z 1=2xcr V2 V1 =V2 V1 1=2
Hence, for downdip shooting, the refraction apparent angle of dip in the direction of the
traveltime function is given by: shootingline (xaxis direction).
In case of one horizontal refractor, the sufciently accurate; provided that the
refraction traveltime function, is given by: refractordip is less than about 10 (dip < 10).
h i1=2
Ta x=V2 2z V2 2 V1 2 =V1 V2
5.2.6 Refraction in a Multilayer
Medium
That is:
dTS dTR T0 =2
slope = 1/V1
surface
Z1
V1
Z2 V2 > V1
Z3
ic V3 > V2 ic
V4 > V3
function of depth, V(z) = a + bz, (a and b are consist of two parallel but displaced linear seg
constants), then the traveltime function is given ments of slope equal to the inverse of the velocity of
by Nettleton (1940, pp. 257261) and Dobrin the faulted layer (slope = 1/V2). These two seg
(1960, pp. 7778): ments correspond to the refracted arrivals from the
reflectors original surface and from the down
Tx 2=bsinh1 bx=2a thrown surface respectively.
In reference to Fig. 5.15, the fault throw (Z)
The raypath, in this case, is of circular shape of is calculated from the difference in intercept
radius (r = zm + a/b), where (zm) is the maximum times (T), and the angle (ic) which is equal to
depth of penetration which is given by: cos1(P/Z). Since P = T V1, and (sin
h i1=2 ic = V1/V2), we can write:
zm x=22 a=b2 a=b h i1=2
DZ DT V1 V2 = V2 2 V1 2
Geometrical shape of raypaths and traveltime
curve are shown in Fig. 5.14. For more accurate determination of the throw
(Z) and fault location, another reversed refrac
tion survey is needed to be conducted.
5.2.8 Refraction from a Faulted
Refractor
5.2.9 Applications of Seismic Wave
Refraction surveying can be used to detect a fault Refraction
and determine its throw. Consider a twolayer
system where the highvelocity layer (velocity, V2) Refraction eld work involves laying out a
is vertically faulted (fault throw, Z) as shown in spread made up of sources and detector, that suite
Fig. 5.15. The traveltime curve of refraction recording both of the direct and refracted
experiment across the strike of the fault plane will wavearrivals. Because of the relatively long
5.2 Seismic Wave Refraction 101
distance
0 x
c1 c2 c3 line of centres
r1 a/b
earth surface
zm x
0
surface
Z1
ic Z2
fault throw, Z i P
c V1
V2
travelpath, refraction arrivals suffer from atten (ii) Smallscale refraction surveys
uation in both of the energy and high frequency (iii) Seismic fan shooting
contents. To lessen these adverse effects, an (iv) Seismic tomography
adequate increase in energy source and adequate
(i) LargeScale Refraction Proling
detector frequency response are used.
Largescale refraction can be used in exploring
The main elds of application of refraction
geological rock layering of regional dimensions.
waves are:
Sourcereceiver spreads can be employed to
(i) Largescale refraction proling
survey lines of several kilometers to several tens
102 5 Seismic Wave Transmission and Refraction
A
source
points
B
5.2 Seismic Wave Refraction 103
sourcepoint
receiverpoint
against sourcereceiver distances. For more the surface. Like any other inversion process,
accurate mapping of the anomalous body more seismic tomography aims at determination, as
than one source point is used for the same accurately as possible, of the earth structural
receiver array. This is similar to the scheme used model for the area under study.
by the modern tomography procedure followed Seismic transmission tomography deals with
in exploring anomalies by sourcereceiver seismic transmission waves which, in general,
recording systems. More detailed description of experience refraction processes at boundaries
the method is presented in the geophysical liter where velocity changes. There are two types of
ature like Nettleton (1940), Bth (1971) and seismic tomography, reflection and transmission.
Kearey et al. (2002, p. 116). Reflection tomography deals with reflection
traveltime measurements, while transmission
(iv) Seismic Tomography
tomography deals with sourcetoreceiver
The word (tomography) is of Greeklanguage
raypaths along which the seismic waves are
origin, meaning (crosssection drawing). In the
directly transmitted. The more commonly
seismic eld, tomography found its way as an
applied technique is the transmission,
effective exploration tool under the name of
boreholetoborehole (also called crosshole)
Seismic Tomography. This technique has been
technique. In this method (crosshole method), a
applied successfully in imaging seismic velocity
set of source points are distributed down a
variation and producing subsurface geological
borehole and another set of receivers are located
models. Seismic tomography is considered to be
in the other borehole, as shown in Fig. 5.17.
one of the modern techniques used in seismic
Another form of the seismic tomography is
exploration.
the surfacetoborehole technique. In this method
Seismic tomography is a type of inverse
the source points are located on the surface and
modeling (inversion process) which uses
the receiver points are distributed down the
acquired seismic data to generate earth models
borehole, as shown in Fig. 5.18.
(usually in terms of velocity and dimension
Interpretation of the data recorded from the
information). The more applied technique in this
crosshole surveying (or from that of the
eld is the (transmission tomography) which
surfaceborehole) aims at velocity distribution
requires putting the source points in a borehole
in the medium between the source and receiver
and the receiver points at another borehole or at
104 5 Seismic Wave Transmission and Refraction
sourcepoint
receiverpoint
sets. Iterative, forward modeling is usually cess, until computed traveltimes become
applied with the help of an especiallydesigned closest to the measured traveltimes. The end
tomographyalgorithm, starting with an objective of the whole process is to determine
assumed model of velocity distribution. The the subsurface geological model as accurately
model is changed after each comparison pro as possible.
2D Seismic Reflection Surveying
6
The seismic reflection method depends on the produces a data volume in which the seismic
same principle of the wellknown echo phe amplitude is function of the variables (x, y, & z).
nomenon. A mechanical shock generates a seis Reflected waves carry useful exploration
mic wave (normally, Pwave) which propagates information in two forms: the traveltime and
through the earth material. On hitting a surface of waveform changes. Both of these two types of
a rock layer (an interface), the incident changes are functions of the physical properties
waveenergy is partly transmitted (passing and layering geometry of the subsurface geology.
through the interface), and partly reflected back In other words, the seismic information furnished
to the earth free surface. by reflection surveying, can supply two main types
The earliest successful application of seismic of exploration knowledge which are: structural
reflection technique was made about 1930 (Net exploration information from traveltime mea
tleton 1940, p. 233). Since then, and due to its surements, and stratigraphic exploration informa
great efciency in the exploration work, the tion from waveform measurements. These two
method was intensively applied in exploring types of changes constitute the principal seismic
subsurface oiltraps. In its conventional form, the tools for exploring the subsurface structural and
technique is based on initiating of a seismic wave stratigraphic geology (Fig. 6.1).
and recording its reflection arrivals by detectors
deployed on the earth surface along a straight
line. The normally applied technique, the seismic 6.1 Reflection Surveying Concepts
reflection surveying, is based upon the use of a
linear sourcereceiver spread moving along a 6.1.1 The Spread Configuration
dened seismic line. To cover a complete survey
of an area, the process is repeated for all lines of The shooting spread used in seismic surveying is
a dened linenetwork designed for that area. dened as the geometrical relationship between
The end result of this technique is obtaining a the location of the sourcepoint and that of the
twodimensional seismic section for each seismic receiverarray. This shotreceiver conguration is
line, which is, in effect, expressing the variation kept unchanged during surveying of the seismic
of seismicreflection amplitude as function of the line. The main elements of a spread are the
two variables: distance, x and depth, z. The sourcepoint (commonly known as the
section lies in the vertical plane that coincides shotpoint), and number and interspacing of the
with the surface surveyline. For this reason, the receiverpoints. In practice shotpoints and
method is called twodimensional 2D surveying receiverpoints are not points but sets of points.
to differentiate it from the 3D surveying which Thus a receiver point is usually made up of a
STRUCTURAL STRATIGRAPHIC
EXPLORATION EXPLORATION
group of detectors (geophones) and the shotpoint multiple of the number (12). In 2D seismic sur
consists of a group of sourcepoints. That is why veying the number of spreadchannels used has
these are commonly referred to as geophone increased from 48 in the early 1970s to 240; and
group and shot pattern. For computation pur 360 channels in the following forty years. In 3D
poses, their geometrical centres are used to rep surveying the number may exceed 1200.
resent the shotpoint and receiverpoints locations
respectively. The shape of the shooting spread and
6.1.2 The Seismic Trace
its main elements are shown in Fig. 6.2.
Based on the position of the shotpoint in
The seismic trace, considered to be the building
relation to the receiverpoint, there are three types
brick for the whole seismic survey data, is a graph
of spread conguration in common use. The
representing the recorded amplitude variation as
shotpoint may be located at one end of the
function of reflection traveltime. The trace con
receiver linear array giving a type called (endon
tains all the outputs of an active seismic channel
spread). Another type of spread is the one in which
during the recording time. It is, in fact, a record of
the shotpoint is located within the spread it is
the whole sequence of the reflection arrivals as well
called (split or unbalancedspread), and when it is
as other arrivals such as direct waves, refracted
located exactly at the centre of the receiver
head waves in addition to the various types of
arrayline, it is called (centre spread). The three
seismic and noneseismic noises. There are three
common spread types are shown in Fig. 6.3.
common modes of tracedisplay. These are:
The number of the active receiver points (or
seismic channels) per spread is customarily xed (i) Wigglytrace, when the amplitude variation
at a wholenumber which is customarily taken as is represented by a wiggly line recorded.
Shotpoint
(i) endon spread
Shotpoint
(ii) unbalanced spread
Shotpoint
(iii) centre spread
(ii) Variablearea trace, when the peakparts reflection point (RP) or common depth point
of the wiggly trace are blacked in. (CDP). Referring to Fig. 6.5, a sevenchannel
(iii) Variabledensity trace, when the ampli spread (endon type) produced a shot record
tude variation is represented by an appro containing seven traces that have been reflected
priate shadeintensity. from corresponding seven CDPs.
An actual record of a 48trace shotpoint
Usually, the eld shotrecords are displayed in
obtained from a centrespread is given herebelow
wigglytrace mode (Fig. 6.4).
Fig. 6.6. It is showing fairly clearly, reflection
Seismic stack sections are displayed by super
arrivals as well as direct and refracted arrivals.
imposing the wiggly and variablearea modes of
display.
6.1.4 The Seismic Profiling
6.1.3 The Shotpoint Technique
A seismic survey of a certain seismic line starts Seismic surveying along a linear track is carried
with deployment of the spread elements (shot out by a technique known as (multichannel
and receivers) along that line. The designed proling technique). The survey procedure
spread, which consists of linear array of involves the use of a xedshape spread that
receiverpoints which are colinear with the moves along a linear course at a regular moveup
shotpoint, is laid down along the line to be shift. The rst shotrecord is obtained from the
surveyed. When activating the energy source (as spread which is laid out at the startpoint of the
ring a dynamite charge), a seismic wavefront line. In the following step, the second shot is
advances through the medium in the form of recorded after shifting the spread location by a
spherical surfaces (spherical when the medium is certain stepup distance. This process
homogenous and isotropic). The seismic rays are (shiftandshoot process) is repeated from start
reflected from intervening interfaces, back to the point to the end point of the line. The moveup
surface to be detected by the geophone groups distance is normally made to be integral multiple
planted on the surface. The reflection arrivals are of the receiverpoint spacing. A map of all
then fed into the recording system via the elec locations that have been occupied in surveying
trical channels which are displayed in a the entire line is called (surfacecoverage map).
tracegather (the shotgather). For more clarication of the proling tech
Each recorded trace (provided by an active nique, an example is given in Fig. 6.7 which
channel of the spread) represents a series of wave shows a seismic line of length of 39 receiver
arrivals of signals reflected from a subsurface stations, surveyed by an endon spread. The
108 6 2D Seismic Reflection Surveying
Fig. 6.4 Part of 48channel shotrecord showing reflection arrivals and random seismic noise. Traces are displayed in
wigglytrace mode
surface coverage, shown in Fig. 6.7, consists of midpoint. This means that each receiverpoint
seven positions of the spread that advanced along of a spread will receive a seismic arrival reflected
the line at moveup shifts of four from a CDP located below the sourcereceiver
receiverstations per shift. midpoint. In fact, shooting of one spread, the
Because this procedure gives a seismic section reflector will be sampled by a number of reflec
representing a prole of the earth along the sur tion points (CDPs) equal to the number of the
veyed seismic line, it is referred to as a proling spread active receivers. From the geometry of the
technique. spread raypath, the CDP spacing will be half
that of the receiver points of the spread. Thus, the
subsurface coverage (CDPline) covered from
6.1.5 The Fold of Coverage shooting one shot will be about half the spread
length (Fig. 6.8).
By denition the CDP is a subsurface point Seismic proling procedure is accomplished
located vertically below the sourcereceiver by repeating the shooting process (shot ring and
6.1 Reflection Surveying Concepts 109
R1 R2 R3 R4 R5 R6 R7
S
Reflector1
CDP1 CDP3 CDP5 CDP7
TX
Reflector1
Reflector2
Reflector3
reflection record
Fig. 6.5 Shotpoint raypaths and its corresponding seismic shot record. Spread (endon type), made up of 7
receiverpoints (R1, , R7) producing 7 seismic traces (TR1, , TR7) reflected from 7 CDPs (CDP1, , CDP7)
reflectionarrivals recording) at a series of equally it is shown that when the spread moveup is too
spaced shotpoint locations along the surface line. large (larger than half the spread), the subsurface
By making the spread moveup distance equal to coverage is incomplete, with gaps in the CDP
half of the spread length, one reflection per one sequence where patches of the reflector are giv
CDP will result for the entire line. The resulting ing noreflection information Fig. 6.9a. When the
seismic section in this case is commonly known 12channel spread is made to advance by
as (singlefold section). For larger moveups, 6receiver station moveup rate, the reflector is
gaps will result in the sampled reflector. That is, completely covered with singlefold coverage
certain CDPs will not have reflections at all. On (Fig. 6.9b). In the third case (Fig. 6.9c), the
the other hand, if the moveup distance is smaller spread is made to move with a spread moveup
than half the spread length, then CDPs will get distance of two receiverstations, which resulted
more than one reflection, giving the case of in threefold coverage for the whole line except
multifold proling technique. The number of for the lessthan threefold zones (called fold tails)
reflections realized per CDP is called the fold of at the beginning and at the end of the line. In this
coverage, (or just, fold). Of course the shorter the example, the fold, in the tailzone, is building up
moveup, the greater the fold will be. from a single fold at the rst CDP to threefold
Considering an example of a seismic proling over 8 CDPs. The same thing occurs at the end of
case in which a twelvechannel spread is used. the line but in reverse direction.
The effect of the moveup distance on the fold of In a regular shooting, which is normally
coverage is claried in Fig. 6.9. In this example, applied in seismic proling, the fold of
110 6 2D Seismic Reflection Surveying
Fig. 6.6 An actual 48trace shot record showing clear reflection arrivals with relatively highamplitude direct and
refracted arrivals. Traces are displayed in variablearea mode of display
coverage (F) can be calculated from the fol there is a direct relationship between fold and
lowing formula: reflection signalquality appearing in the stack
trace. However the relation is not linear but fol
F N=2n lows a power law of the form (F1/2) (Robinson
1983a, p. 111). Accordingly, the rate of change
where, (N) is the total number of the spread in the (S/N) enhancement diminishes with the
active receiverchannels and (n) is the moveup increase of fold (Fig. 6.10).
distance expressed in number of receiverstation
spacing covered in one moveup distance.
This equation shows that the fold is directly 6.1.6 The CDP Trace Gather
proportional to the number of channels and
inversely proportional with the moveup distance. As mentioned above, a shotgather is the group of
The fold of coverage is a major factor in traces recorded by the spread channels of that shot.
increasing the signaltonoise ratio (S/N). In fact Thus, a sevenchannel shot has seven recorded
6.1 Reflection Surveying Concepts 111
receiver stations
1 10 20 30
source receivers
S1
S2
S3
S4
S5
S6
S7
Fig. 6.7 Surface coverage of a line made up of 7 spread moveup shifts
source
receivers
Fig. 6.8 Shooting spread surfacecoverage, reflection raypath, and CDP subsurface coverage, as realized from one
seismic shotpoint
seismic traces. Likewise, the CDPgather is a this regrouping process (trace resorting), a simple
group of seismic traces which belong to one proling example is given here (Fig. 6.11).
CDP. Naturally, the number of traces in a certain In this example (shown in Fig. 6.11) we have
CDP is equal to the fold of coverage in that CDP. a survey made up of ve shots (A, B, C, D, and
In a certain seismic proling process, the traces E). ShotA of traces (a1, a2, , a6) and shotB of
obtained for each shot (shotgather), are regrouped traces (b1, b2, , b6), and so on, for the rest of
into CDP gathers ready for being stacked after shots. The complete surface coverage is occu
being subjected to certain processing steps which pying a total of 11 receiverstations with one
aim at enhancing the reflection signal. To explain receiverstation for the spread moveup distance.
112 6 2D Seismic Reflection Surveying
(a)
moveup distance1 moveup distance2 moveup distance3
S1 S2 S3 S4
(c)
Moveup distance1
Moveup distance2
Moveup distance3
Moveup distance4
S1 surface coverage
S2
S3
S4
sub surface coverage
CDPs
Fig. 6.9 Surface and subsurface coverage in seismic Spread move distance is 2 receiverstations, resulting in
proling surveying using an endon spread. Moveup nominal fold of 3, with tailfold at beginning and end of
distance: greater than half spread length (a), equal to half the line
spread length (b), and more than half spread length (c).
Corresponding to this surface coverage, is a known the (Coverage Diagram). In Fig. 6.11, the
subsurface coverage made up of 14 live CDPs, shotgather is indicated by the horizontal arrow, and
(CDP1, CDP2, , CDP14). The trace gather for the CDPgather is indicated by the vertical arrow.
CDP1 (CDPG1) consists of one trace (a1). For Number of traces in each gather, represents the fold
the second CDP (CDPG2) the gather has also of coverage. In this example, the fold is equal to 3
one trace (a2). The CDPG3 has two traces for the range (CDP5CDP10), and tail fold for the
(a3 + b1), and so on. The number of traces per ranges (CDP1CDP4), and for (CDP11CDP14).
CDP is representing the fold of coverage. The redistribution of the traces from the
Identifying the trace gathers (for shots and shotgathers to CDPgathers (called CDP sorting) is
CDPs) is much simplied by drawing what is shown in the following gure (Fig. 6.12).
6.2 Seismic Reflection Data Acquisition 113
Fig. 6.10 Curve expressing type of dependence of the 6.2.1 Seismic Energy Sources
S/N ratio on fold of coverage of reflection signal. The
curve is of the powerlaw form (F1/2)
The traditional method applied in generating
seismic waves is exploding dynamite in shot
holes. There are however, other methods which
6.2 Seismic Reflection Data have been introduced as alternative seismic energy
Acquisition sources. Choice of the sourcetype depends on the
surface conditions prevailing in the given survey
In seismic reflection data acquisition, we are, in
area. The main criteria considered in evaluating a
general, dealing with generating and detecting
particular sourcetype are the following:
Pwaves which are propagating from the source
region and getting reflected from subsurface strong enough to generate strong seismic
interfaces. The complete data acquisition process signal that can be reflected from deep inter
involves seismic source generation, reflected face and be detectable at the farthest receiver
wave detection, and digitally data recording. (receiver of maximum offset).
In order to record the reflection arrivals with the generated seismic signal is rich in high
minimum distortions, optimum operation parame frequency components to be able to resolve
ters (called eld parameters) are determined and closely spaced reflectors.
applied in conducting the seismic survey. The the generated noise is of least energy level.
SHOTS receiverstations
& 1 2 3 4 5 6 7 8 9 10 11
CDPS
FOLD 1 1 2 2 3 3 3 3 3 3 2 2 1 1
Fig. 6.11 Example of proling process, where a horizontal arrow and the CDPgather is indicated by the
6channel endon spread moving at one receiverstation vertical arrow. Nominal fold is 3 over the central 6 CDPs
moveup distance. Shotgather is indicated by the and the tailfold zones are at beginning and end of line
114 6 2D Seismic Reflection Surveying
LVL
weathering zone
dynamite charge
seismic rays
surface (low velocity layer, LVL), the high fre the bottom of a heavy vehicle to increase coupling
quency components are severely attenuated. This of its plate with ground surface (Fig. 6.15).
method is rarely applied at present. On detonating the gas mixture, a sudden
pressure impact occurs which is transmitted
(iii) Gas Exploding
through the baseplate to the ground surface. By
This is another impulsive source (known as
hydraulic system the plate is locked into position
Dinoseis) in which the energy is created by
after the impact, to prevent repeated impacts.
exploding a mixture of oxygen and propane gases
Being a relatively weakenergy source, more
contained in a conned chamber, the bottom of
than three units are red simultaneously by a
which is a moveable plate resting on the ground
control signal sent from the recording system,
surface. The so designed chamber is attached to
and the shot is repeated many times, then vertical
stacking is applied to get improved S/N ratio.
The general features of Dinoseis are similar to
the Weight Drop method. Both are surface
sources which are generating relatively weak
falling mass seismic energy, developing strong surface waves,
and producing lowfrequency seismic pulses.
The method is rarely applied these days.
Fig. 6.14 WeightDropping seismicenergy source
gas chamber
116 6 2D Seismic Reflection Surveying
(iv) Dynamite Cords Exploding in laying the explosive cords. Since the method
This seismic source, well known by the name depends on trenches for the explosive cords, it
(Geoflex), consists of an explosive cord which is means that it is not possible to be used in areas of
buried in the ground at a shallow depth (about too hard surface layer or in cities or villages. It is
half a meter depth). It is laid down by a not advisable to be used in areas where the sur
hydraulicallyoperated plough which is espe face layer is too thick, since in such conditions the
cially designed for this purpose (Fig. 6.16). transmitted energy will be much weakened.
Implementation of the Geoflex method starts
(v) Vibroseis
with laying the explosive cord (about 100 m
This is a nonimpulsive seismicenergy source
length) inside a trench of about a halfmeter
which was introduced in early 1950s and rapidly
depth and detonated at one end. While the
gained popularity as an acquisition tool. In 1982,
explosivegenerated energy is travelling very fast
over 40 % of worldwide seismic surveys were
through the cord from the detonation point,
using it (McQuillin et al. 1984, p. 38). Unlike
seismic waves are generated from each point of
impulsive sources, Vibroseis method creates
the cord at delayed times. This will cause these
mechanical energy which is continuously vibrat
waves to have phase differences bringing about
ing for certain time duration. A mechanical, truck
the ltering effect similar to that made by
mounted vibrator is hydraulically driven, produce
sourcearrays. Points of the earth surface coin
electronically controlled vibration. The generated
ciding with the cord, are receiving the liberated
energy is conveyed to the ground by a metal pad
energy not at the same time, but in sequencial
(about 1 m2) attached beneath the truck
manner. For this reason, the downgoing
(Fig. 6.17).
wavefront becomes inclined with respect to the
When in operation, the Vibroseis system
horizontal level. Along this wavefront (shown
transmits into the earth a seismic signal vibrating
by dotted lines in Fig. 6.16), highest energy level
at frequency which is varying linearly with time.
is obtained. In effect, there will be a ltering
This electronicallycontrolled vibration function
effect such that the nearsurface horizontally
is called the Sweep. In practical application, the
travelling energy (surface waves) will be ltered
sweep duration (called sweep length) is normally
out before reaching detectors.
within the range (1020) s. Over this timespan
The main advantages of the Geoflex method, is
the sweepfrequency applied is (1050) Hz.
the attenuation of the surface waves, in addition
Apart from the taper imposed at both ends of the
to generating wideband seismic waves. Regard
sweep (about half a second in length), the
ing eld operation the method is fast and simple
explosive cord
Detonation
point wavefronts
of
generated
Seismic waves
6.2 Seismic Reflection Data Acquisition 117
vibrating
metal pad
amplitude
time
(c)
taper zone
amplitude
time
118 6 2D Seismic Reflection Surveying
In order to extract the impulsivelike record of is the (Unipulse), where the air continues
the reflections in their actual reflectiontimes, the flowing into the generated bubble for some
recorded traces are crosscorrelated with the time after the initial discharge. This is done
applied sweep. This process can be done in the to lessen the effect of the sudden bubble
eld within the recording unit of the Vibroseis collapse which results in bubble oscillation.
system, or later by the processing software (see (ii) The SteamGun (also called Vaporchoc)
processing section). utilizes hot steam instead of air. It is
The main advantages of the Vibroseis method considered to be better than the AirGun,
are being fast, safe, and comparatively cheap to because with using hot steam, the effect of
run. It can be applied along roads and even in cities bubble oscillation is reduced.
since it causes no damaging effects on environ (iii) The GasGun is still another similar marine
ments. Technically, it has the advantage of having source which uses explosion of a mixture of
the sourcefunction being under control. The propane and oxygen contained in a steel
sweep parameters (time duration, frequency range, chamber. The pressure pulse created by the
and taper) can be changed at will. The main explosion of the propaneoxygen mixture is
problem with the method is the low source passed to the water via a flexible seal which
energylevel. Like other weak surface sources, forms part of the chamber. The (Aquapulse)
vertical stacking is carried out at the same time as works on this principle.
the recording is going on. In normal surveying (iv) Dynamite explosion is implemented as a
work, several Vibroseistrucks (typically 4) are marine energy source. A method called
operating at the same time, and about 2060 (Maxipulse) uses small dynamite charges
vibrations are conducted per each shotpoint red directly in the water body. Another
location. The shotgather traces per each shot, similar method (Aquaseis) uses long det
are obtained from vertical stacking of all of the onating cord.
recorded traces involved in the one shotpoint. (v) Electrical Arc Discharge is used to gen
erate energy by electrical discharge in
Marine Seismic Sources water. Examples of this method are the
The most common types of seismic energy (Sparker) and the Wasp. This is a rela
sources are: tively weak energy source which is more
(i) The AirGun is a widely used energy source suited for shallowgeology exploration.
in marine seismic surveying. It generates Out of these methods, the AirGun is by far the
energy by discharging highly compressed most commonly used type of energy source in
air into the water. A variation of this method
6.2 Seismic Reflection Data Acquisition 119
seismic marine surveying. In certain situations, the land and hydrophones for use in marine
Air Gun, is used as a land seismic source by gen environments.
erating a seismic pulse from an AirGun sub
(i) The Geophone
merged in a waterlled pit made on the earth
Detection of reflection arrivals in land surveying
surface. Typical example of such an application is
is done by geophones (seismometers as they are
in providing the seismic energy for VSP surveying.
sometimes called). The geophone is a
It should be noted here that the names of the
seismicdetection instrument which can trans
energy types are trademarks given to them by
form the ground vibration motion into an elec
their originators. The following Table 6.1 con
trical voltage. The most commonly applied
tains a summary of the types of seismic sources,
geophones in seismic reflection surveys are the
with their trademark references.
electromagnetic type. It operates on the principle
of voltage generation in a coil moving within a
6.2.2 Seismic Detectors magnetic eld. The generated voltage is propor
tional to the velocity of the motion of the coil
Earth movement due to the arrival of a seismic relative to the magnet. For this reason, geo
wave is generally very small. The vibration phones of this type are normally referred to as
amplitude is of the order of 108 of an inch (velocity geophones).
(Gardner 1938). In addition to the The geophone is made up of a permanent
smallamplitude feature, the seismic waves have cylindrical magnet and a coil suspended by
two other criteria which the detection and leafsprings inside a circular slit made in the mag
recording equipments should be able to cope net. The slit is separating the magnetic southpole
with. These are the wide amplituderange and the (inner part of the magnet) from the northpole
wide frequencyrange. (outer part). The magnet is rmly attached to a case
The seismic detection process is based on which is tted with a spike for easy xing on the
conversion of the ground vibration to an electri ground in an upright position (Fig. 6.20).
cal signal by a special transducer which can For operation, the geophone is planted verti
respond to seismic amplitude and frequency full cally into the ground. In this position the magnet
ranges and without distortions. Seismic detectors will vibrate in vertical direction when the ground
are of two main types, the geophones for use in is seismically activated, and the coil stays
Fig. 6.20 Schematic damping maximum damping applied without stopping the
representation of the resistor oscillation. Damping of about 0.7 times the
electromagnetic geo
phone. Symbols N and critical damping gives a response which increa
S denote magnetic north N S N ses smoothly with increase of frequency. This
pole and south pole damping factor (D = 0.7, in Fig. 6.21) is the
respectively damping which is normally applied in seismic
surveying environments.
(ii) The Hydrophone
This is a geophoneequivalent detector used in
magnet detecting seismic waves in marine surveying
environments. The hydrophone is a
pressuresensitive detection device that uses
spike
substances that generate electrical voltage pro
portional to pressure changes caused by the arri
val of a seismic signal. Such a substance (called
piezoelectric substance) has a property of gener
ating an electric voltage when subjected to pres
sure. Piezoelectric transducers are also called
stationary because of its inertia. The generated electrostrictive devices (Sheriff 2002, p. 263). The
voltage in the coil is function of the vibration rate pressure changes, caused in a water medium, due
and the coil parameters (number of turns, radius, to passage of a seismic wave, are proportional to
and eld intensity of its magnet). the velocity of the water particles set into motion
In order to reduce oscillation, geophones are by the signal (Dobrin and Savit 1988, p. 63).
normally provided with a damping resistor In marine surveying, hydrophones are housed
shunted across the geophone terminals. The in a special hose lled with a certain liquid of
damping device controls the frequencyresponse such a density that makes the net density of the
of the geophone vibration. Typical forms of the hose (containing the hydrophones and liquid)
geophone response characteristic curves for approximately equal to that of water in order to
different damping factors, are shown in facilitate controlling its depth below water sur
Fig. 6.21. face and in reducing effects of external sudden
Too heavy damping (overdamping) reduces forces affecting the streamer (cable jerking).
detection sensitivity, while too little damping Marine data acquisition is conducted by tow
(underdamping) will lead to continued oscilla ing the hose, equipped with hydrophones (called
tion. Critical damping is dened to be the streamer) behind the survey boat (Fig. 6.22).
frequency
Natural
frequency
6.2 Seismic Reflection Data Acquisition 121
energy source
6.2.3 The Seismic Data Recording signal. With this unit the toosmall amplitude is
increased and the toolarge amplitude is reduced.
For about 30 years after seismic exploration was This operation results in compressing the signal
introduced, seismic shotrecords were directly amplituderange to that value which is matching
recorded on paper as wiggly traces. In the early the dynamic range of the recording unit. The
1950s, however, recording of seismic data on technique of amplitude matching with the
magnetic tapes was introduced in analogue recorder, for undistorted output, is an essential
mode. After about ten years, the more superior feature of all seismic recording systems. A mod
digital recording technique was introduced. Soon ication of the AGC is the programmed gain
after mid 1965, the analogue method was com control (PGC) with which the gain is not tied up
pletely superseded by the digital method. Despite to the input amplitudevariation. A PGCcoupled
the fact that analogue recording is now consid amplier has its gain that varies with record time
ered to be obsolete, (that is, no more in appli by a preset (programmed) timedependant func
cation), a brief description of the analogue tion. This is usually made available as an option,
system is presented herebelow. This will be applied when it is required.
useful as the digital system involves some of the Next to the amplication part is the magnetic
components of the analogue system. recording system which produces a permanent
record of the data stored on magnetic tapes. This
(i) The Analogue Recording System
is accomplished either directly or after being
The analogue recording system carries out the
frequencymodulated (FM). A carrier frequency
recording in two stages: the amplication process
(typically, 4 kHz) is modulated by the signal
followed by the recording process. In the rst
received from the amplication system. In this
stage, the seismic signal, received from the
way it can produce a recorded signal of S/N ratio
geophone group, is amplied so that it will be at
up to 60 db, and frequency response which is
a level that suits the following recording stage.
practically flat over 3200 Hz (Evenden et al.
The seismic signal may be too small for the
1971).
recorder to detect, or too large for the recording
Through a playback of the recorded magnetic
unit to handle without being saturated. In order to
tape, or directly from the amplier output, a
cope with this wide range of amplitude variation
visual display may be obtained. The display unit
(estimated to be 80100 db) amplication factor
(camera unit) is usually included in the recording
is made to be variable. This is done by using a
system. This unit uses the principle of gal
special electronic unit called the automatic gain
vanometer deflection caused by voltage variation.
control (AGC), which makes the amplier gain
By certain optical arrangement, galvanometer
to vary with the amplitude level of the incoming
122 6 2D Seismic Reflection Surveying
magnetic tape
AGC camera recorder
geophone
group
AMP FIL
AGC
AMP FIL GC
DA
Fig. 6.24 Schematic diagram showing the signal flow in a digital recording system
d5 d4 d3 d2 d1
D
c5 c4 c3 c2 c1
C
b5 b4 b3 b2 b1 .. d3 c3 b3 a3 d2 c2 b2 a2 d1 c1 b1 a1
B Sample1 Sample2 Sample1
a5 a4 a3 a2 a1
A
multiplexing
Fig. 6.25 Principle of the multiplexer operation, tracesequential input with samplesequential output, using 4
channels (A, B, C, D)
To conceive the multiplexing operation, let us d1) in this sequence. In the second rotation cycle
consider a mechanical device in the form of an arm the outputted amplitudevalues (samples) will be
rotating with constant rotation speed where the (a2, b2, c2, and d2). With continued rotation, the
arm head will trace a circular ring. Referring to output through a single channel will continue
Fig. 6.25, consider four channels, arranged feeding the AD conversion unit, until the end of
equally spaced on that ring. In the rst rotation the recording time. In this way, the seismic data
cycle, the arm shall make contacts with the chan has been converted from tracesequential
nels (A, B, C, and D) in that order. In this cycle, the amplitudevalues to samplesequential mode, as
multiplexer will output amplitudes, (a1, b1, c1, and it is shown in this gure (Fig. 6.25).
124 6 2D Seismic Reflection Surveying
twenty years ago, the tapes in common use were 21 When the cell is magnetized it represents (1) and
trackone inch tapes or 9 trackhalf inch tapes. In when it is not, then it is (0). This concept is claried
that recording system the rst track is usually from the following example (Fig. 6.26).
assigned for the parity check. At present narrower An analogue signal, shown at the top of
tapes (e.g. 4 mm tapes) and higher recording Fig. (6.26), is digitized and its sample values are
density tapes (in forms of small cartridges) were measured and converted into 8bit binary words.
introduced for the seismic data recording. The eight bits of each sample (each binary word)
are recorded across the tape by magnetizing the
(iii) Concept of the Dynamic Range
cell when the bit is (one) and demagnetizing the
Any measuring or sensing instrument (geophone,
cell when it is (zero). The sign bit (marked as
amplier, recording system) has a bounded
Sbit) is a flag indicating the algebraic sign of the
capability of faithful detection. The system
sample value. It is assigned the value of (one) for
noiselevel sets the detection lower limit, while
negative sign and (zero) for positive sign. For
the system technical characteristics set the
example, a sample value of minus nineteen
detection upper limit. For a given system, dy
(19), the binary word in this example (9track
namic range is dened to be the ratio (normally
tape) is (100010011), and for a value of (+27),
expressed in dbs) of maximum measured
say, will be (000011011), where the rst digit is
amplitude to the minimum recoverable signal,
for the sign (the sign bit).
where the minimum signal is taken to be the
In the case the recorded data is in multiplexed
noise amplitudelevel.
form, the sequence followed will be in sample
The dynamic range (DR) is expressed by the
sequential, and if the data is in demultiplexed
ratio:
form, the recording will be in tracesequential
DR A=adb mode. Normally, the magnetic tape is divided
into blocks where the rst block (called the
where, (A) is the maximum signal amplitude header block) carries general information like
which a system can measure (with tolerable record number, sampling period, record length,
distortion level) and (a) is the system noiselevel. and other recording parameters. The next block is
The concept is often applied in evaluating the the seismic data which are usually in multiplexed
detection and measuring capabilities of recording form. Along with each sample value, the applied
systems. The early seismic analogue systems gain for that sample and signbit flag are recor
which were recording directly on paper have ded. One track is reserved for the paritycheck
dynamic range of about 25 db. The dynamic for checking the bitcolumn across the tape. The
range of analogue systems recording on magnetic end block contains information signaling the end
tapes is about 45 db. In case of digital recording of that set of recorded data.
systems, the dynamic range is dependent on the The Society of Exploration Geophysicists
number of bits used in expressing the sample (SEG) has developed several seismic data
value. For 14bit recording, it is equal to 84 db. recording formats since the 1960s when the
digital recording system was introduced. In 1967,
(iv) The Magnetic Tape Recording the formats SEGA, SEGB, and SEGX, were
The magnetic tape is a plastic strip, coated with published. Later on, other formats were pub
magnetic material, is divided into tracks of cells. lished, as SEGC in 1972, SEGY in 1975 and
After being digitized, each samplevalue is repre SEGD in 1980. The formats of SEGC and
sented by a binary word which is recorded on the SEGY are recording the data in demultiplexed
magnetic tape by magnetization status of the cells. mode, whereas SEGD can record in both
126 6 2D Seismic Reflection Surveying
analogue signal
(+)
0
()
sample values
6 9 22 69 66 15 0 7 4 30 32 11
0
2 0 1 0 1 0 1 0 1 0 0 0 1
21 1 0 1 0 1 1 0 1 0 1 0 1
22 1 0 1 1 0 1 0 1 1 1 0 0
3
2 0 1 0 0 0 1 0 0 0 1 0 1
24 0 0 1 0 0 0 0 0 0 1 0 0
5
2 0 0 0 0 0 0 0 0 0 0 1 0
6 0
2 0 0 0 1 1 0 0 0 0 0 0
2 7 0 0 0 0 0 0 0 0 0 0 0 0
S 0 0 0 0 0 0 0 1 1 0 0 0
digital magnetic tape
Fig. 6.26 Simplied sketch showing the principle of 9track tape recording format. The binary word for each sample
value consists of 8 bits plus the sign bit (S)
multiplexed and demultiplexed modes. At pre (DA) converter, demultiplexer, and the camera
sent, eld recording is mostly done using unit which displays the shot tracegather (usually
demultiplexed SEGD format. in wigglymode traces) on the computer monitor
Developments in the taperecording tech or on paper. A typical shot record is presented in
niques concerned mainly the packing density and Fig. 6.27.
storage capacity. The density was developed This gure shows about one second of data of
from 800, 1600, then 6250 byte per inch. The a shot record made up of 48 seismic channels
tapes used nowadays are of 8 or 4 mm width, plus two auxiliary channels (at the left side of the
contained in cartridges or cassettes. record): the time break indicating the zerotime
of the shot, and upholetime. Traces are dis
played in wiggly mode of display.
6.2.4 Data Playback and Display
For quality control (QC) purposes, the end result 6.3 Seismic Noise Characterization
of the recording station (the shot tracegather) and Attenuation
needs to be displayed. The playback process is
the reverse of the recording process. It is done As we mentioned in a previous discussion, all
through a complete playback system which nonreflection arrivals are considered noise which
consists of a digital AGC, a digitaltoanalogue cause distortions and masking effects to the
6.3 Seismic Noise Characterization and Attenuation 127
+ + +
receiver
output
receiver array
reflection wavelets
reflector
Fig. 6.28 Role of a multidetector receiver in attenuating random (and coherent) noises and enhancing reflection signal
of a vertically arriving reflected wave
a group of shot holes of a shotpoint pattern. source) is used. The array sourceelements
Compared with a singleelement receiver, a (shotholes, for example) are detonated simulta
multielement receiver (receiver made up of neously generating body waves which reach the
many geophones), has the capability of attenu detector array after being reflected from subsur
ating horizontally moving waves in addition to face interfaces. At the same time, the
enhancing the signaltorandom noise ratio by a sourcegenerated surface waves travel from the
factor depending on the square root of the source to the receiver. As it is with the receiver
number of the array elements. Efciency of the array, the nearly vertical incident body waves are
array in attenuating surface waves is expressed in phase and hence they interfere constructively
by a characteristic function called the array to give enhanced reflected signal. For attenuation
response, dened to be the ratio of the amplitude of the surface waves, the length of the source
of an arrayoutput to that of the same number of array should be equal to the wavelength of the
the arrayelements gathered together at one generated surface waves.
location (Sheriff and Geldart 1995, p. 247). This The response of any given array (having given
is expressed in the form of a mathematical number of elements with given element spacing)
function called (directional response function, or can be presented in the form of a curve known as
just directivity function). (response curve or directivity curve). The
The theory, on which the concept is based, is abscissa is the ratio of the wavelength of surface
the same whether it is applied to the shot waves (coherent noise) to element spacing, and
holepattern or to the receiver geophone array. the ordinate is the array output expressed as the
Instead of using a single source point, a source ratio of the arrayoutput to that of the same
array (shothole pattern, in case of dynamite number of elements gathered together at one
6.3 Seismic Noise Characterization and Attenuation 129
location. The array response is normally emphasized that (k) is representing the apparent
expressed in db units. wavelength measured in the array direction.
The basic principle is that waves travelling in Typical response curves (directivity curves) of
near vertical direction are enhanced while those geophone lineararrays are shown in Fig. 6.29.
travelling horizontally are attenuated. This tech The characteristics of the array response are
nique has been borrowed from engineering work summarized as follows:
done in radioantenna design (Dobrin and Savit
The response curve repeats itself at
1988, p. 99). The application of the principles in
(x/k = 1), that is at (x = k), so it is com
geophone and sourcearrays was published as
pletely dened in the interval (x/k = 01).
early as mid1950s (Lombardi 1955) and (Parr
The curve is symmetrical about the point
and Mayne 1955).
(x = k/2)
(ii) The Array Response Function It has (n) lobes (maxima), where the principal
Directivity curves express the array response to main lobe centered at (k = ) and another
the different wavelengths of horizontally travel equal lobe (the rst alias lobe) centered at
ling surface waves. For a linear array consisting (k = x)
of n elements with x interspacing, the response It has (n 1) nodes (zeroresponse) at
function, R(b), is given by: k = nx, nx/2, nx/3, , n x/(n 1). At
these points the attenuation is innite, that is
Rb sin nb=sin b there is no output.
It has (n 2) side lobes found in the range
where, b = px/k, and k is the apparent wave k = nx to nx/(n 11). This zone, called
length of the surface waves travelling in the the (reject zone) or the (effective array
direction of the linear array. length), is the zone where the most effective
The R(b) function is periodic, repeating at a noiseattenuation occurs. Sometimes the
period of (b = p). The function is fully dened in reject zone is taken to be the zone found
the range of (b = 0 to p), that is, in the range between the 3 or 6 db points. A response
(x/k = 0 to 1). The curve, within this range, is curve for an 8 elements linear array is shown
symmetrical about the point (b = p/2), that is in Fig. 6.30.
about (x/k = 1/2), and it crosses the baxis at
This gure shows that, the attenuationzone
the points b = p/n, 2p/n, 3p/n, , (n 1)p/n.
(reject zone) lies between (k = 8x) and
The curve has a total of (n) maxima, where the
(k = 8x/7). This zone can be made wider (ex
principal one is located at (b = 0), and the peaks
tended) and its response level lowered further, by
of the side lobes occur at the mid points between
increasing the number of elements (n). However
the zeros of the function.
this will cause narrowing of the pass band
Mathematical derivation is based on summing
(marked by the rst lobe where k nx) and
of spatial harmonic waves moving along a linear
increasing its cutoff slope.
array made up of n elements, arranged at con
The role of the directivity curves is in evalu
stant interspacing x (Sheriff and Geldart 1995,
ation of array performances in attenuation of
p. 247).
horizontally moving surface waves. Given the
(iii) The Array Response Curve wavelength of the dominant coherent noise, a
The response curve is normally plotted as relative seismic detector array can be designed such that
response against (x/k), or against (k). The its response is most effective in attenuation of the
response value is plotted directly or in db units. noise having that wavelength. From the drawn
As it is mentioned above, the response, by de curve, one can directly read the relative
nition, is expressed by the ratio between the array noisecancelation effect corresponding to the
output and that of the same elements gathered at wavelength of the dominant surface waves (k), or
one location, that is when (x = 0). It should be to the parameter (x/k). This means that we need
130 6 2D Seismic Reflection Surveying
response
2geophone linear array
array
0.5
0.5
0.5
Fig. 6.29 Schematic representation of response curves for different linear arrays
1.0
main alias
lobe reject zone lobe
0.5
R()
x/
0
0 1/8 2/8 3/8 4/8 5/8 6/8 7/8 1
Fig. 6.30 Schematic representation of response curve for 8element linear array, showing the main and alias lobes in
addition to the arrays reject zone
6.3 Seismic Noise Characterization and Attenuation 131
to know the wavelength of the surfacewaves in The xed spread method is done by using
order to design the geophone array which is endon spread which is made up of geophones
appropriate for the attenuation of the surface spaced at short distances (typically, 5m spacing).
waves. The wavelength, which is representative After the rst shot is red and the shot record is
of this type of coherent noise, is normally produced, the shot is moved along the spread line
determined by special seismic experiment called by a distance equal to the spread length, and the
the (noise test). second shotrecord is obtained. This
shiftandrecord action is repeated a number of
times such that the resulting combined record will
6.4 Field Measures for Signal have its maximum offset to be about equal to the
Enhancement maximum offset planned for that survey. An
alternative method (called, the xedshot method)
Before commencing a seismic reflection survey, is conducted by xing the shot location and move
certain eld procedures are normally carried out the receiver spread (along the spread line) away,
to determine the nature of the dominant noise in hence, the term (walkaway test).
the area in order to be able to design the source In both methods, each receiver point is
and receiver parameters which can output stron occupied by a single geophone or by a group of
gest reflection signal and least possible noise geophones gathered in one location to give
level. For noise analysis, a special seismic clearest noise possible (Fig. 6.32).
experiment is done in the eld, called the noise The more commonly applied method is the
test (called also, walkaway or microspread xedspread method because it is easier to move
test). Analysis of the resulting seismic record the shot location than moving the spread.
will lead to determination of coherent noise Noisetest operation is normally including
parameters such as apparent wavelength, veloc measures for studying the broadside noise as
ity, periods, and frequency, which are necessary well as the inline coherent noise. For this pur
in the design process for both of the source and pose, another spread is laid perpendicular to the
receiver arrays. normal inline spread (Fig. 6.33).
The end result of the eld operations in con
ducting the noise test is a noisetest record in which
6.4.1 The Noise Test the 5mspaced traces are drawn. It is important to
note that no ltering should be applied during the
There are two types of eld procedures that can recording process. For interpretation purposes, the
be followed to conduct the noise test in the eld. noiserecord display should be produced without
These are the xedspread method and the any type of timevariant scaling. Any amplitude
xedshot method (Fig. 6.31). distorting process should be avoided in order to be
shot
receiver points
Shotpoint
crossline
receiver
spread
geophones
grouped in one
location
Fig. 6.33 Noise test Lshaped spread to detect both of the inline and crossline coherent noises. A receiver point may
be occupied by a single geophone or by a group of geophones
able to measure relative amplitudes, and for easier Interpretation of the noisetest data is nor
event identication. However, timeinvariant scal mally done visually. It starts with identication
ing can be applied. Principal seismic noise events of the types of noises and other types of seismic
(and other wave arrivals) are schematically shown events. The main events a noisetest record
in Fig. 6.34. contains are: direct waves, refracted waves,
offset
A
B F
E
C
F
C
D time
noise cone
Fig. 6.34 Schematic representations of the main events, e backscattered waves, f reflected waves. The group of
normally found in a noisetest record. a Direct wave, events between events (a) and (d) forms the noise cone
b refracted wave, c ground roll, d air wave,
6.4 Field Measures for Signal Enhancement 133
630m/s
2270m/s
reflection event
970m/s
450m/s
350m/s
Fig. 6.35 Actual noisetest record, showing reflection, refraction, sound, and surface waves
ground roll, airwave (sound wave travelling period (s) are manually measured from the paper
through air), backscattered waves, and wave record. From these parameters, frequency (f), and
arrivals of energy reflected from deep interfaces. wavelengths (k) are then calculated using the
The strongest events, appearing in a noisetest relations (f = 1/s, k = v/f). Interpretation results
record, consist of surface waves, covering a are found to be as quoted in the following
range of apparent velocities and apparent wave Table 6.2.
lengths. These events make up the coherent noise These data show that the most distortive event
forming a group of largeamplitude events is the surface wave3 (of velocity 970 m/s and
known as the (noise cone). An actual record of wavelength 108 m). It has the largest amplitude
noise test is shown in Fig. 6.35. and largest extent, covering all the record time
In interpreting the actual noise record, given and all of the offset range in this exampler. On
in Fig. 6.35, it is found that it contains body average the wavelength (100 m) can be consid
waves (direct, refraction, and reflection events), ered as representative of the surface waves in the
surface waves (ground roll), and air wave (sound survey area, and can be adopted in the design of
wave). The wave parameters: velocity (v) and the geophone and shothole arrays.
6.4.2 The Experimental Shooting done visually to determine the best parameters.
Choosing the appropriate parameter is based on
The source parameters depend on the type of the reflectionsignal strength, frequency content,
source used to generate the seismic energy. For the reflectionsignal resolution, noise level, and
dynamite source, there are three main parameters. extent of penetration depth.
These are: charge depth, charge weight, and
shothole pattern. The optimum charge depth and
charge weight are determined through direct 6.4.3 Determination of the LVL
experimental shooting. For the charge depth, Properties
several trial shots are conducted such that the
charge weight is xed while varying the depth. The earth surface is characterized by its
The same approach is followed in optimizing the nonuniform topography (variable elevation),
charge weight that is by xing the depth and and the surface layer is, in general, made up of
varying the weight of the charge Fig. 6.36. lose lowvelocity material. This layer, (com
As for the shothole pattern, the suitable monly referred to as the lowvelocity layer, LVL)
number of shotholes is determined by applying is madeup of one or more layers of velocity and
the same principles applied in the case of the thickness that can vary with location within the
receiver geophone array. However, very often survey area. The LVL thickness and velocity are
and for economical reasons, this is determined by typically of ranges (1050) m and (5001500)
the experimentalshooting method alongside with m/sec respectively.
the determination of optimum charge depth and Because of the large velocity contrast nor
charge weight. Conducting multishot sources is mally found between the LVL material and that
another approach for increasing the of the medium below it, the base of the LVL acts
signaltonoise ratio. This is normally done in as a strong reflector and refractor. Multiple
case of surfacesources as in the case of vibroseis reflections (reverberation and ghosts) can also
and weight dropping techniques. A typical plan develop in such environments. Effect of the LVL
for an experimental shooting for optimizing the is not restricted to the traveltime changes, but
source parameters is as shown in Table 6.3. also on the reflection waveform. In particular,
The shot records for these trial shots should be highfrequency components of the travelling
outputted with no ltering and with no seismic waves experience severe attenuation in
timevariant scaling being applied to them. the LVL due to the relatively strong absorption
A comparative study of the obtained records is phenomenon.
Fig. 6.36 Experimental shooting for determination of: a charge depth and b charge weight
6.4 Field Measures for Signal Enhancement 135
Table 6.3 A typical plan for an experimental shooting for optimizing the source parameters
No. Shothole depth (m) Number of holes Charge weight (kg)/hole Shothole Spacing (m)
1 6 2 3 25.0
2 6 3 3 12.5
3 6 4 3 8.3
4 9 2 4 25.0
5 9 3 3 12.5
6 9 4 4 8.3
7 12 2 3 25.0
8 12 3 3 12.5
9 12 4 4 8.3
An integral part of the eld activity in a points arranged at certain depths inside the
seismic reflection survey is determination of the borehole Fig. 6.37.
properties of the surface layer, the LVL. In par These dynamite capsules are red in sequence
ticular, effort is made to determine the thickness starting at the base of the hole and continuing
and velocity which are essential information upward till the last shot which is nearest to sur
needed in correctioncomputations of the reflec face. For recording the arrivals at the surface, a
tion and refraction travel times. The commonly group of geophones are usually planted at equal
applied methods are the uphole surveying and distances from the surface location of the hole.
specially designed refraction surveying. For analysis and interpretation of the recorded
data, travel times (reduced to vertical travelpath)
(i) LVL Characterization by Uphole
are plotted against depth. From the slope of the
Surveying
resulting plot, and the depths of the points at
The uphole survey involves drilling a borehole
which slopes show abrupt changes, the velocity
of depth exceeding the expected LVL thickness,
and thickness of the LVL layer (or layers) are
normally within the range (50100 m). Small
calculated.
charges (dynamite capsules) are red at a series of
(a) (b)
shot hole detector vertical time, Tcos
depth LVL
(h) thickness
slant
shot time, T
depth
(h)
Fig. 6.37 Uphole survey: a travelpath and b plot of vertical travel time (Tcosh) against depth (h), where (T) is the
slant time and (h) is the angle between travelpath and the borehole
136 6 2D Seismic Reflection Surveying
(a) (b)
time time time
distance distance
direct arrival
refracted arrival
Fig. 6.38 Two methods of refraction survey specially designed for determination of the LVL properties (thickness and
velocity), a xed spread, and b xed shotpoint method
6.5 The Seismic Field Crew 137
(ii) Topographic Surveying Section responsible for the dynamite storage and
Fixing on the ground, of the survey points transport taking all the safety and secu
(shots and receivers) and measuring coor rity precautions.
dinates (x, y, & z) of each of these points. (v) Mechanical Engineering Section
These data, in addition to data concerning For mechanical work needed by the crew
geographical nature and surface environ as maintenance of drilling machines,
ments, are documented and reported. trucks, electricity generators and the
(iii) Drilling Section Vibroseis systems.
This is concerned with drilling the (vi) Administration and Finance Section
shotholes, and other holes needed by the This section is responsible for personnel
survey, such as the deep holes needed for recruitment, living requirements, trans
uphole surveys. port, communications, material storage
(iv) Shooting Section and all nance affairs.
Workers in this section do all the nec
These are the main sections of a typical crew
essary steps needed to prepare a
using dynamite for the seismic energy source. All
shothole. This involves preparation of
the sections are headed by the party chief who is
the right amount of charge and placing it
managing the crew work through direct contacts
at the required depth and lling the
with the crew and through the daily meetings
hole with water and mud mixture to
held every night with the sections heads.
secure coupling. The section is also
3D Seismic Reflection Surveying
7
In the years following the dramatic introduction threedimensional shape, which is spherical, in a
of the analogue and digital data recording and homogeneous medium. In reality, the geological
processing, the main developmenttrends were in medium is madeup of different rock layers of
increasing recording channels and in more different physical properties and different geo
advanced magnetic tape technology. These metrical shapes. With this realistic type of envi
developments have led, in the early 1970s, to ronments, the advancing wavefront, is still
more effective exploration techniques including having the three dimensional form but no more
the introduction of seismic three dimensional perfectly spherical.
(3D) surveying. The rst proper 3D exploration The created seismic rays travel in every
work was conducted using xed cross spreads by direction in the space surrounding the source
Walton in the early 1970s (Walton 1972), and zone. In other words, the geological eld, as well
soon after the 3D seismic method entered appli as the seismic eld in which it is created, is 3D in
cation on commercialproduction basis. At pre nature (Fig. 7.1).
sent the 3D seismic surveying became very In the conventional 2D shooting, the seismic
common especially for detail exploration and eld is created as threedimensional waveeld,
development of oil elds. but detected by twodimensional array of
detectionpoints. Thus, the 2D surveying is
concerned with a limited portion of the reflected
7.1 Introduction 3D waveenergy. The rest of the reflected energy,
with all the useful information it is carrying, is
7.1.1 Nature Is Three Dimensional left to pass undetected.
In the 3D technique the source energy is uti
The subsurface geology targeted by seismic lized more efciently, since most of the reflected
exploration is, in essence, threedimensional in 3D wave front will be detected. It is, in fact,
nature. The seismic eld created in seismic considered to be the more logical approach for
reflection surveys is likewise, threedimensional seismic exploration than the conventional 2D,
(3D). In an ideal homogeneous medium, the because the 3D surveying conforms to the 3D
wavefront of the advancing seismic wave is of nature of the subsurface geology.
(a) Subsurface geology is 3D in nature of the transittime with the depth of the well
which is also a 1D function.
In the conventional 2D seismic surveys,
where the shotpoint and receiverpoints are
colinear, the resulting stack section consists of
a series of seismic traces each of which belongs to
a CMP location. The stack traces of a stack section
are uniformly spaced along the distance coordi
nate (x). Thus the seismic amplitude in the pro
(b) Seismic field is 3D in nature duced section is function of both traceposition
(x) on the seismic line and the twoway reflection
time (t). This means that the amplitude in the
produced seismic section, is represented by the
twodimensional (2D) function f(x, t), or f(x, z)
when time is scaled by the propagation velocity.
With the 3D shooting technique, the
produced CMPlocations form a twodimensional
array over the surveyed reflectors. In this case, we
Fig. 7.1 By nature, both of the subsurface geology have a seismic trace for each of these CMPs,
(a) and the seismic eld (b) created to explore it, are three
forming a data volume in which the seismic
dimensional
amplitude is represented as a function of its
7.1.2 1D2D3D Terminology position in space, dened by the three
dimensions (x, y, and t) or by the dimensions (x,
A seismic trace can be viewed as a time function y, z) when the third dimension is expressed in
of the seismic amplitude which is normally rep terms of depth, z. Thus, the amplitude is expressed
resenting variation of vibrationvelocity as as threedimensional (3D) function, f(x, y, t) or
function of recording time. If we disregard its f(x, y, z).
position information (that is neglecting the xy In reference to (Fig. 7.2), the three terminol
coordinates of its CMP location), the seismic ogy denitions are considering the seismic
trace is represented as a onedimensional (1D) amplitude (a) as function of depth (z), function of
function of amplitude with time, f(t), or with (x, z), or function of (x, y, z).
depth, f(z). The synthetic seismogram is another The end product of the 3D seismic surveying
example of the 1D seismicfunction. It represents is a seismic data volume which represents a
reflectionamplitude variation with the depth of threedimensional function a(x, y, z) expressing
the drillhole. The sonic log expresses variation the variation of seismic amplitude (a) with the
1D 2D 3D
y
x
Fig. 7.2 Seismic data represented as 1Dfunction (seismic trace), as 2Dfunction (seismic section), or as 3Dfunction
(seismic data volume). The corresponding functions are a(z), a(x, z), a(x, y, z)
7.1 Introduction 141
three coordinates: x, y, and z. In this way, one (i) Erroneous Image Positioning
may consider the 2D section as a special case of Geometry of the 2D seismic line, assumes that
the more general 3D datavolume, in which one the sourcereceiver line, the incident seismic ray,
of the dimensions (x or y) is equal to zero. In the and the reflected ray are all lying in one vertical
geophysical literature it is customary to use the plane (the raypath plane). Further, the reflection
terminology (databox, or datavolume) for the point is located vertically below the
whole stackdata set outputted by the 3D survey. sourcereceiver midpoint. This is, in fact true
For display purposes, 2D sections from the data only when the reflector is a horizontal plane
volume, in both vertical and horizontal direc surface or the seismic line is shot along the
tions, can be obtained. nonpinging axis of an anticline for example
(Fig. 7.3a). In the case of a dipping reflector,
however, the reflection events may be received
7.1.3 Limitations of 2D Seismic from points located outside the vertical plane
Surveying which is, in 2D surveying, assumed to be the
plane of the reflection raypath (Fig. 7.3b).
Despite its outstanding success in subsurface A dry well may be obtained due to erroneous
exploration which has been achieved throughout positioning of the seismic image that may appear
the past years, the conventional 2D seismic on the 2D seismic section.
method is facing certain difculties in retaining An example showing erroneous image
its wellestablished success standard. This is due positioning is the case of a seismic line passing
to the fact that most of the large structural traps nearby a dome. The 2D line does not cross the
with clear geophysical anomalies have already structural dome and yet the produced seismic
been discovered and what remained undiscov section shows an anticlinal image (Fig. 7.4).
ered are those which are characterized by being Appearance of the image of a subsurface
small in size, complex in structure, weak in dome, which is not crossed by the vertical plane
seismic response, and may be situated in inac containing the seismic line, is due to reflections
cessible areas. Limitations of the 2D seismic from the sides of the offside dome.c
surveying may be summarized as follows:
(ii) Distortion due to Dipping Reflectors
Strictly speaking, the 2D surveys give true struc
A B tural pictures only when the surveyed areas are
made up by horizontally layered geology. In
RCVR line general, there is always a certain amount of
image distortion produced by 2D method when
ever dipping reflectors exist. In processing of 2D
data, the reflection events are assumed to have
been received from reflection points which are
CMP line
located vertically below the surface seismicline.
When the reflectors are not horizontal, and not
continuous, planes the 2D data will give erroneous
dipping subsurface images. With 2D migration, distortions
reflectors due to reflectors, dipping in the sourcereceiver
direction, can be corrected, whereas distortions
Fig. 7.3 Reflection raypaths for zerooffset receivers in due to dipping in other directions cannot be cor
2D surveying. a Line shot along strike direction, located
directly above an anticline of nonpinging axis. b Line
rected. Regarding dipeffect on seismic results we
shot parallel to line (a), but at horizontallyshifted position can distinguish the following three cases:
142 7 3D Seismic Reflection Surveying
2D seismic line
time
2D seismic section
Fig. 7.4 2D seismic line shot and an offside dome. The produced seismic section shows a false structure due to
reflections from the side of the offside dome
R
S R S R
RP RP
RP
(a) (b) (c)
Fig. 7.5 Distortion effect of dip direction on reflection point location. a Case of no dip, b dip in the SR direction, and
c dip is perpendicular to SR direction. S and R are the source and receiver points respectively
(Case1) Dip is of zero value (Fig. 7.5a) (Case3) Dip has component perpendicular to
In this case, where the reflectors are the shotreceiver direction (Fig. 7.5c)
horizontal planes, the reflection point is The reflection point, in this case, is
located vertically below the shot shifted outside the shotreceiver ver
receiver midpoint, hence the reflec tical plane. In this type of situation 3D
tion images, produced by 2D data, are survey has to be conducted and the
found in their proper positions, and no resulting datavolume must be sub
distortions occur in this case. jected to 3D migration process in
(Case2) Dip is in the shotreceiver direction order to get the correct subsurface
(Fig. 7.5b) structural image.
In this case, the reflection point is Distortion always occurs in case of dipping
shifted updip by an amount of shift reflectors and the distortionseverity depends on
which is dependent on the dip value. the dip and on dip direction. Thus, if a line in
The 2D method is adequate for this type 2D survey happened to be along the
of situation provided that the resulting dipdirection, then the reflection raypath will be
stack section is corrected through in the sourcereceiver vertical plane, and in this
an appropriate 2Dmigration process. case, no outofplane reflection points shall
7.1 Introduction 143
occur. The dipcaused distortion in this case, is On migration of the sections of these two
correctable in the processing stage (by applying intersecting lines, the strikeline section experi
2D migration). In the general case, where the 2D ences no change whereas all events on the
line does not exactly coincide with the true dip dipline section will be updip shifted. It is
direction, the reflectionevents will not be cor expected that this situation will lead to a mistie
rectly repositioned even with the process of 2D on the intersection point of the two sections. In
migration. In fact the amount of correction will fact, a stack section of a strikeline has two
be corresponding to that component of the dip problems. These are: creation of a mistie (at the
found along the line direction. To get the com intersectionpoint) with the section of the dip
plete migration, the data must be subjected to line, and the misallocations of the reflection
another migrationprocess to be implemented in a CMPs which are not lying in the vertical plane as
direction perpendicular to the line direction the 2D stack sections are normally displaying.c
which is, of course, not possible. It can be con
(iv) Weak Resolution of Structural Changes
cluded, therefore, that the 3D data volume,
Due to the wide spacing of the 2D lines, con
obtained from a 3D survey, is a necessary
struction of the isochron contour maps are based
requirement in order to achieve proper migration
on interpolation of the reflection times measured
process regardless of the dip direction of the
on the survey lines. For this reason, structural
subsurface reflectors.c
details are generally not resolvable by the 2D
(iii) Creation of Misties surveying. Structural resolution gets worse as the
Let us consider the zerooffset sections (as stack linespacing gets large compared with the sizes
sections) of two intersecting seismic lines of the structural anomalies. The isochrones map
(dipline and strikeline) shot over a dipping of an area covered with ve 2D lines is shown in
plane reflector (Fig. 7.6). (Fig. 7.7a). In this case, the 2D survey failed to
For the strike line, the reflection raypath resolve the three domes detected by the 3D sur
plane is slant and not vertical, as it appears in the vey carried out in the same area (Fig. 7.7b).
seismic stack section, and the reflection events The weak resolution power of 2D surveying
obtained from the dipping reflector shows no becomes more serious problem when the struc
dipping feature. On the other hand, the dip line tural anomalies are of dimensions smaller than
will be of slant raypaths, all lying in the vertical the applied linespacing.
raypath plane.
7.1.4 Merits of the 3D Technique
dip line
surface plane
Due to economic constraints the 3D method is
strike line used only in cases when more detailed and more
zerooffset
reflection accurate exploration are aimed at. Typically, it is
raypath
applied for restricted purposes such as develop
ment and appraisal of already discovered
zerooffset oilelds. When economy allows, however, 3D
reflection CMPs
raypath P surveying is applied in the normal general
Reflection
purpose seismic exploration.
dipping plane As we have seen above, the 2D method gives
CMPs
distorted images of the subsurface geology espe
cially in case of dipping layers. The 2D method also
suffers from weakness in its ability of resolving
Fig. 7.6 Effect of dipping reflection plane on zerooffset
small geophysical anomalies and ne structural
reflection points in case of two 2D intersecting lines.
Strike stack section shows no dip, whereas dip line does. and stratigraphic details. Inability of surveying
On migration, a mistie occurs at intersection point (P) inaccessible areas is yet another limitation of
144 7 3D Seismic Reflection Surveying
Line5
the 2D seismic technique. These limitations were direction, in addition to possibility of getting
removed or very much reduced by the application horizontal sections (timeslices) at any level
of the 3D method. The advantages of the 3D sur within the datavolume (Fig. 7.8).
veying may be grouped under the following three
(ii) Capability of Surveying Inaccessible
groups of application domains:
Areas
(i) In the Geophysical Domain As far as seismic surveying is concerned, an area
3D surveying is furnishing good control on the is considered to be inaccessible, when neither
determination of processing parameters, such as seismic energysources nor receivers are allowed
static values, stacking velocity, and dipvectors, to be located within the area boundary. An
and hence it allows more effective three inaccessible area can be surveyed by deploying
dimensional migration process. Increased ef the receivers on the boundary of the area and
ciency in the use of seismic source energy, due to shooting at sourcepoints which are also dis
the large number of active detection channels tributed over the area boundary. The surveying is
normally used in 3D eld acquisition. The 3D data conducted by shooting the source points in
processing provides the stack datavolume, which sequence while the receiver spread is kept xed
allows displaying vertical sections at any throughout the shooting process. For each shot,
3D data box
vertical
ysections
vertical vertical
xsections sections
horizontal
sections
7.1 Introduction 145
source
point
there will be a number of CMPs located at the (iii) In the Geological and Reservoir
sourcetoreceiver midpoints, which will be Domains
located within the surveyed area (Fig. 7.9). The 3D data can furnish accurate information on
By this type of shooting (called loop the subsurface geological structure of the area.
cshooting), the area will be covered with CMPs Unlike 2D data, it can resolve small and complex
without having survey points (sources or recei structural and stratigraphic anomalies. It provides
vers) existing within the area. Special software direct information (with no interpolation proce
can handle the recorded data and get it processed dure) of the subsurface geology expressed in
to produce a 3D datavolume. its real threedimensional image. Due to the
Although this technique makes seismic sur dense sampling points of the geological space,
veying of an isolated inaccessible area possible, the 3D data becomes more readilyinterpretable,
it suffers from a number of weak points. The leading to increased degree of accuracy and
distribution of both of the recorded CMPs and much improved resolutionpower (Fig. 7.10).
the fold of coverage are not uniform. Also, it is In the reservoir domain, 3D data provide
possible to lose shallow reflections with traces delineation of the 3D shape and spatial limits of
recorded at largeoffset receivers. These limita oilelds as well as giving more accurate esti
tions become less effective, the smaller the area mation of their hydrocarbon reserves. The data
and the deeper the targeted reflectors. Applica can help in outlining oil/water contact which can
tion of the concept (loop spread surveing) was assist in more accurate emplacement of devel
dealt with in more detail in (Alsadi 1992, 1994). opment and production wells. Reservoir studies,
based on 3D data, help in more accurate reservoir study and prepare its nal report for guiding work
characterization in terms of facies distribution in the plan implementation.
and fluid content. Such information would lead The preplanning phase is an important pre
to more accurate oileld characterization and requisite needed in order to facilitate the work
enhanced production rates. execution especially in the dataacquisition
stage. Preknowledge of the area assist the geo
(iv) In the Economical Domain
physicist to be prepared, and make the necessary
Despite the overall high cost and long survey
provisions for any problem that may crop up
duration compared with the 2D method, the 3D
during the work. Examples of such problems are:
method is considered to be costeffective when the
skip shots, makeup shots, seismicenergy to be
merits of its results are taken into consideration.
used, work permit, and other administration and
From the economical point of view, 3D data would
logistical matters.
give increasing successratio (producertototal
The preplanning report of a 3D seismic survey
wells ratio) and improving well productionrates.
normally covers the following basic items:
The overall eld evaluationcost is reduced
through less developmentwells, proper well (i) Compilation and assessment of available
emplacement and shorter developmentperiod. technical data.
In view of its exploration efciency, the 3D (ii) Survey objectives.
method gives an enhanced success ratio of (iii) Operational constraints.
development drilling and increase in the oil (iv) Surveydesign plan.
reserves. These features form the basis for the (v) Scouting and preliminary tests.
justication of conducting 3D surveys especially (vi) Work organization and scheduling.
for oileld appraisal and eld development. In (vii) Cost analysis.
comparison with development drilling, the justi (viii) Health, safety, and environment
cation is even stronger in comparison with the protection.
2D method. For example, a 10 km2 3D survey
It should be noted here, that the item of the
may cost no more than 10 % of that of drilling a
contract preparation is deliberately omitted from
well of 3000 mdepth. The overall saving
the above mentioned list. The omission of this item
achieved from a 3D survey is represented by
is made on the basis that it represents an indepen
reduced eldevaluation costs and by earlier
dent piece of work to be carried out in the following
production due to shortened developmentperiod.
stage, rather than being part of the preplanning
report. This is natural procedure, since the contract
items are depending on the survey parameters
7.1.5 Survey Preplanning which need to be prepared and made ready before
starting of contract drafting. Herebelow we shall
Although, in principle, work preplanning is give brief explanatory notes for each of the items
essential in any seismic surveying project, it included in the preplanning report.
becomes more necessary prerequisite when deal
ing with 3D surveying. The urge for the preplan
ning phase stems from the complex nature of the 7.1.6 Cost Considerations of 3D
3D surveying work which involves large nancial Surveys
investments. A feasibility study should be con
ducted as a joint effort by a team of personnel from 3D surveying is usually quoted as being a costly
all parties concerned with the 3D survey project. In process. To be more objective, however, the cost
particular, the team should include specialists in the assessment must be made in relation to the
data acquisition, data processing and interpretation, turnover benets. In fact, even 2D surveying can
as well as experts on legal and contractual matters. become very expensive when using very small
The task of this team is to carry out the preplanning line spacing, and high fold of coverage. It is
7.1 Introduction 147
generally accepted now that the 2D and 3D the region of (1015) % of the cost of the data
methods become equivalent costwise when the acquisition. These gures are nearrealistic when
2D lines are spaced by about half a kilometer. the surveypoints for both methods are spaced by
The 3D surveying is considered to be 50 m and the coverage folds are 1600 and 3200 %
costeffective compared with the 2D method for the 3D and 2D respectively. In general, the
considering the merits of the 3D data. Compar cost of 3D surveying depends on the surface
ison between 2D and 3D surveycost is more environment of the survey area (Fig. 7.11).
meaningful, when it is realized that 3D data
(ii) Dependence of Cost on Survey
provide dense spatial sampling of reflectors,
Parameters
enhanced structural resolution, and undistorted
The survey parameters which are most influential
3D images.
on survey cost are the spread parameters, fold of
The 3DSurvey Cost Elements coverage, bin size, and shot density used in a given
In discussing survey costs, and 3Dto2D cost survey. These elements are interrelated in their cost
comparison, it is of course very difcult to quote contribution. Thus, fold (F) may be statistically
precise gures. However, it is possible to give a estimated from the formula:c
fair assessment of the main elements of the cost.
These elements are important issues to be con F Ntr =Nbin
sidered in cost analysis of any 3D surveying
project. where, Ntr and Nbin are the survey total number
The main costelements of the 3D survey can of traces and total number of bins respectively.
be summarized as follows: By substituting the product (shot density,
SHD) x (survey Area) x (live channels per shot,
(i) Surface Conditions NCHN), for Ntr, and (survey area/Bin Size) for
Difculty in quoting precise gures is due to the Nbin, we get the important relation:
wide variation between types of surveys. In par
ticular, prevailing surface conditions, and types of F SHD NCHN Bin Size
the applied survey parameters differ from survey
to survey. In the present worldwide seismic Hence,
industry, the cost of the dataacquisition of
onshore surveys (onland surveys) is approxi SHD F=NCHN Bin Size
mately in the range of (510) k$/km for 2D
method and (2040) k$/km2 for 3D. The cost of Since total surveywork depends on shot
processing and interpretation is estimated to be in density (SHD), this formula is showing that the
total work done in conducting a 3D survey (and
cost hence total cost) is directly proportional to fold of
k$/km2 coverage and inversely proportional to bin size
and number of active channels used per shot.
100 A closely related to cost, is the (turnaround
time), or the duration of the complete 3D survey.
For a given survey (given area size and given
50 survey parameters), survey duration depends on
the workproduction rate, which is, in turn,
farm, shallow Jungle dependant on the applied work parameters, like
offshore desert urban water survey parameters, type of energy source, surface
conditions, eld processing, topographic survey,
Fig. 7.11 Sketch showing dependence of survey cost permit fees, and number of workshifts fol
(cost of eld data acquisition) on surface environment lowed in conducting the survey.
148 7 3D Seismic Reflection Surveying
It is found that the eld dataacquisition of a subsurface circle will have a radius of half that of
land 3D survey of a 500 km2area, say, would the surface circle, and all reflected waves received
last for about (23) months. Adding to this, time by the detectors, distributed over the surface cir
needed for processing and interpretation, the total cle, will have arrived at the same time. However, if
duration time to complete a 3D project of such an the receiver points are laid down on a straight line
area is estimated to be in the order of about (68) passing through the source point, the reflection
months. points from which the incident energy is reflected
and detected by the colinear receivers, will fall on
a subsurface straightline with pointspacing equal
7.2 Definitions and Basic Principles to half that of the receive points. In this case, the
detected reflection information (seismic ampli
7.2.1 Definition of 3D Surveying tude) is restricted to the 2D vertical plane con
taining the reflaction raypaths (Fig. 7.12).
Consider a case where seismic energy, generated As this gure is showing, the difference
at a surface point, is reflected from a subsurface between 2D survey and 3D survey is based on
horizontal planer reflector. The generated seismic the way the spread elements are distributed on
wavefront, which is of 3D form, spreads out the surface. In fact, what makes a survey 2D or
through the geological medium (assumed to be 3D is the way the receivers are deployed on the
homogenous) with a velocity decided by the surface. In 2D surveying, receivers are laid down
physical properties of the medium. At the instant in line with the source point, and the used spread
the wavefront hits the reflection plane, a circular is linear, made up of onedimensional array of
zone of it (of the reflection plane) will be illu receivers, and in this case, the reflection points
minated by the incident wavefront which is, will fall on a subsurface straight line which is the
according to reflection laws, reflected back to the projection of the receiver line onto the horizontal
surface. reflectionplane. In the case of 3D surveying, the
Using the rayconcept, detection points dis spread consists of receivers which are distributed
tributed over a surface circle centered about the over an area, and the used spread consists of a
source point, will receive reflected energy from twodimensional array of receivers. The reflec
reflection points located on a corresponding sub tion points in this case will be distributed over a
surface circle. Based on the laws of reflection, the subsurface area of the reflectorplane.
RP RP RP RP RP
subsurface reflector plane
7.2 Definitions and Basic Principles 149
The spread in the 2D geometry consists of a Considering horizontal plane reflectors, all the
receiver points colinear with the source point, reflection raypaths, of the CDP tracegather in
whereas the spread in 3D survey takes the geo 2D surveying, are coincident on one common
metrical form of a rectangular grid of receiver vertical plane (the raypath vertical plane). In the
points. Typical 2D and 3D spreads are shown in 3D method, however, each reflection raypath
the following Fig. 7.13. may fall in its own vertical plane. These raypath
Whether it is 2D or 3D survey, the planes are generally not coinciding on each other,
reflectionpoints (RPs), created from ring a because of the varying sourcereceiver directions
shotpoint, are spaced by half the receiver spac (azimuths) of the CDP gathertraces. Thus, the
ing. The spread used for 3D surveying (3D receiver azimuth in 2D surveying is constant. It is
spread) is normally covering a rectangular area. in the direction of the seismic line. The corre
From geometrical consideration, it can be seen sponding azimuth, in 3D surveying, is variable
that the subsurface area covered by the reflection because of varying receiver bearing in relation to
points is also a rectangle which is of an area the source point.
equal to quarter of the area covered by the 3D The raypath planes of the reflection seismic
spread. This can be readily observed from the rays (of a 3Dsurvey shotpoint) are vertical
Fig. 7.13. planes but they assume different directions. This
Shotrecords, obtained from 2D and 3D is the case because the receiver points are located
spreads, are shown schematically in the follow at different bearings in relation to the shot loca
ing Fig. 7.14 tion. The raypath bearing, with respect to the
RP RP RP RP RP
R R R R R
(b) S
RP RP RP RP RP
S RP
source receiver reflection point
150 7 3D Seismic Reflection Surveying
CMPs CMPs
distance, x
time line1 line2 line3 line4
T
shot point, is usually referred to as the azimuth. centered about the CDP. The width and length of
Azimuth variation is considered as an advantage this search area are equal to the CDP spacing in
of the 3D data because with this feature the the two perpendicular directions of the survey
reflection seismic waves shall sample the rock stationgrid. This search area is called the (bin)
medium in threedimensional space and not and the traces included within it form the
restricted to one plane as the 2D data is provid bingather. The process of sorting of traces into
ing. The concept is shown in Fig. 7.13. the appropriate bin is called bingridding or
binning. These concepts are shown in Fig. 7.15.
The bin is normally specied by certain
7.2.4 The CDP Bin and Bin Attributes properties or identication criteria. The bin can be
rectangular or square in shape, having a dened
In normal proling technique followed in 2D dimensions and surface area commonly referred
surveying, the spread moves along a uniformly to as the (bin size). Other parameters of the bin
spaced stationpoints, at equal moveup rate. In concerns its trace gather. These parameters, called
this way, the reflection points of each CDP the (Bin Attributes), include the trace offset, az
tracegather are all coinciding at the imuth, and fold of coverage. Normally, the bin
CDP. Sometimes, due to a shifted placement of a attributes are displayed in a special diagram
shotpoint, one (or more than one) reflection called the spider diagram as shown in Fig. 7.16.
point may fall outside the CDP location. In CDP
sorting, a search distance (usually dened to be
equal to half of the CDPspacing) is set in pro 7.2.5 The Surface and Subsurface
cessing. All traces found within the search dis Coverage
tance of a CDP are included in the trace gather of
that CDP. In the normal swath shooting, the surface of the
The same principle is applied to the CDPgrid survey area becomes covered by two sets of
in case of 3D surveying, except that in 3D, we mutually perpendicular lines (the station lines).
have a (searcharea) instead of the 2D These are the receiver lines and the shotlines.
searchdistance. The search area is normally The corresponding subsurface lines are the sub
having a rectangular shape (or square shape) lines (inlines) and crosslines which intersect at
7.2 Definitions and Basic Principles 151
CDP line
search distance
CDP grid
search area
(BIN)
2 4
3 5 seismic surveying, the corresponding
endproduct is a threedimension space of
trace1 offset azimuth stacked traces called the data volume or the
data box as it is sometimes called. The data
7 6 volume represents a threedimensional function
in which the seismic amplitude varies with the
three coordinates; (x, y, t). The reflectiontime
BIN CDP
(t) may be replaced by the depth dimension (z),
Fig. 7.16 Spider diagram showing the bin attributes where (z = tv/2) and v is velocity.
(offset, azimuth, and fold) of a bin having 7 traces in its The data volume (in digital form) consists of
gather (fold = 7) discrete volumeelements, called the (resolution
cells) dened as parallelepipeds of dimensions
bin centres. By denition, the bin is a subsurface (x, y, t), where (x) and (y) are the CMP
rectangular (or square) area having dimensions spacing in the inlines and crosslines respec
which are equal to half of the corresponding tively. The third dimension (t) is equal to the
surface receiverstation spacing (Fig. 7.17). sampling period of the stack traces. This volume
The subline and crossline seismic sections element is considered to be the building brick of
are made up of sequence of stack traces, each of the data box, the ultimate product of the seismic
which is located at a bin centre. In this sense, 3D survey.
subline and crossline seismic stack sections are The resolution cell is also called the (voxel),
considered as of subsurface locations and not in analogy to the term (pixel) used for the unit of
surface locations. a digital picture. The data volume is a subsurface
dataset, made up of voxels of dimensions (x,
y, t) as illustrated in Fig. 7.18.
7.2.6 The Seismic Data Volume With the help of special processing software,
it is possible to extract a variety of seismic sec
The endproduct of a 2Ddata processing, is the tions from the datavolume. Normally, three
2D seismic stack section. In case of the 3D types of mutually perpendicular sections can be
152 7 3D Seismic Reflection Surveying
sublines
(inlines)
(SL)
crosslines
(XL) XL
CMP bin y
SL
receiver point x
y SL2
x SL1
t
3D data box
(data volume) resolution cell
(voxel)
extracted. A vertical section may be obtained representation of the seismic amplitude values
along any subline, or along any crossline, and a which fall on the same reflectiontime. In fact, it
horizontal section (called a time slice) can be is the locus of equaltime amplitudes existing in
obtained at any reflectiontime. It is also possible the data volume. In color displays, the timeslice
to obtain a vertical section connecting points at plots are given in colorcoded amplitudevalues
arbitrary locations as for example, sections along with clear distinctions between peak and
distances connecting several well locations. troughvalues.
A section which is not a subline or a crossline is The time slice has a similar form as that pre
normally referred to as a crooked, oblique or sented by the time contour map. The difference
diagonal section (Fig. 7.19). between the time slice and the structural contour
The horizontal section, called seiscrop sec map is that in time slice, we display different
tion, or more often called (time slice), is a amplitude values at constant reflection time (time
7.2 Definitions and Basic Principles 153
time slice
of the time slice). The structural contour map and data analyses are applied in the eld to
shows different time values of the same reflection determine the optimum parameters applied in the
surface. Briefly stated we say that time slice is data acquisition.
showing different amplitudes measured at one Here below, are the main aspects of the
timeslice, and in contour map, we display the dataacquisition requirements involved in 3D
same reflection event (belonging to a given seismic reflection surveying. Spread parameters
reflector surface) at different time values. Several and shooting procedure is most important aspect
reflectors may be represented in a given of the 3D eld acquisition activities. Many types
timeslice, but different timevalues of the same of spread are in application, which can be dis
given reflector, are shown in a given contour map. cussed under two main headings: marine and
Derivation of structural contour maps from land 3D spreads.
timeslice sections can, in principle, be obtained
by marking the amplitudes of the same seismic
reflector, on a sequence of timeslices, and pro 7.3.1 Types of Marine 3D Spreads
jecting the picked values for that reflector on one
map. Conventionally, marine 3D surveying is con
ducted by using spreads similar to those used in
land 2D seismic surveying. The boat and the
7.3 3D Field Data Acquisition towedbehind streamer is recording as it moves
along straightline courses. The technique differs
In 2D seismic surveying, the shooting spread from the normal 2D surveying in that the survey
consists of a shot point and receiver points which shooting is conducted along closely spaced par
are all arranged in one straight line (linear allel lines. Line spacing (typically, 100200 m),
spread). In the 3D surveying, however, these is kept constant throughout the survey area.
surveypoints are distributed over an area (areal With the progress that took place in the nav
spread). Essentially the eld work consists of a igation techniques and in navigation data pro
set of procedures which implement certain geo cessing, certain improvements were introduced
physical parameters. The eld parameters are on the applied spreads. In cases of calm seas, the
normally adopted through compromising survey boat and its trailing streamer move along
between the ideal technical requirements and the a straight line. In other cases, and because of
overall acquisition cost. Certain testrecording local sea crosscurrents, the streamer may drift
154 7 3D Seismic Reflection Surveying
source
hydrophone receiver
recorded CMPs
away from the boat linear course. This is referred reflection points varies as shooting progresses.
to as the (feathering effect), and the angle Computations of the reflection points and CMP
between the drifted cable and the linear distributions are done in the processing stage
boattrack is called the feathering angle based on the gathered boat readings of the cable
(Fig. 7.20). geometry during shooting.
The shape of the drifted cable, for each shot, To increase survey efciency, multisource
is determined from the readings of the special and multistreamers have been used in the marine
compasses tted along the cable. Because of surveying. Examples of marine spreads which
changes of sea conditions, the cable shape and can be congured are: single sourcesingle
the feathering angle change from shot to shot. streamer, dual sourcesingle streamer, single
Consequently, distribution of the subsurface sourcedual streamer, and dual sourcetriple
streamer (Fig. 7.21).
Spread movement during recording, is done
(a) normally along parallel straight lines (parallel
linear paths). Because of the long streamerline
towed behind, the boat needs to turn with a rel
(b)
atively large steering radius in order to record the
following line. A variation of this method is
(c) shooting along circular paths. The boat moves
along overlapping or spiral circular paths. The
advantages of the circular type of spread, is that
(d)
the survey is completed with no times lost due to
longtime turns of the boat needed in case of
source
hydrophone receiver straightline surveying. These types of
spreadmovements are shown in Fig. 7.22.
Fig. 7.21 Types of spreads used in marine 3D survey It should be noted here that, deviation of the
ing. a Single sourcesingle streamer, b dual sourcesingle streamer courseline from the linear form (caused
streamer, c single sourcedual streamer, d dual by the circular spread or by the feathering effect),
sourcetriple streamer
Fig. 7.22 Types of spread movements used in marine 3D surveying. a Parallel linear path, b Overlapping circular
path, c Spiral circular path
7.3 3D Field Data Acquisition 155
(d) (e)
Spread
swathmovement along
first spread last spread
Fig. 7.25 Spread movement (shiftandshoot movements) over the swath, with overlapping shooting spreads
7.3.3 The Swath Shooting Technique gives uniform bin fold and adequate offset and
azimuth coverage. On the other hand, swath
The swath is dened to be a strip of the survey shooting requires land with fully accessible space,
area of width equal to the width of the used and type of surface conditions which allow the
spread that moves along the dened strip in a appropriate freedom for the survey maneuvering.
rollalong movement similar to the A typical example, of shooting parameters
shiftandshoot technique used in conventional applied in swath survey, is given herebelow
2D proling surveying (Fig. 7.25). (Table 7.1).
In practice, when the template reaches the end
of the rst swath, the template is recongured at
the nearby end of the adjacent swath, then car 7.3.4 Types of Templates Used
rying out the shiftandshoot process from end in Swath Shooting
to end. In this way all the rest of swaths are
sequentially covered. In general, swaths are not There are many types of templates that can be
laid down side by side, but laid down with used in swath shooting. A template may contain
overlap to obtain fold buildup (Fig. 7.26). one, or more than one, shotlines which are
Greater fold coverage is obtained with bigger arranged to be incentre, offcentre, or at ends of
overlap made between adjacent swaths. the template. The more common types are those
Swath shooting is a common acquisition where the shotlines are perpendicular to the
technique implemented in todays land 3D sur receiver lines. Another variation is to have the
veys. It has the advantages of being simple to shotlines inclined with respect to the
congure, and efcient to execute in the eld. It receiverlines (Fig. 7.27).
swath4
swath3
swath2
1.4 m
array centre 10.0 m
10.0
m
48.56 m
Geophone type/polarity SM4/SEG standard
Natural frequency 10 Hz
Critical damping 70 %
String conguration 12 geophns (2 parallel 6 series)
Receiver array parallelogram
Geophones/geophonegroup 36
Geophone spacing 4.16 m
Number of strings 3
Stagger between strings 1.4
String separation 10 m
Array length 48.56 m
Array width 20 m
(continued)
158 7 3D Seismic Reflection Surveying
5m
array centre
37.5 m
Source type Vibroseis System (vibrator)
Peak force 45,000 lb
Number of vibrators 4
Sweeps per VP 1
Sweep length 16 s
Sweep frequency 864 Hz
Sweep taper 500 ms at both ends
Source array Parallelogram 2 2
Vibrator inline stagger 12.5 m
Vibrator lateral spacing 5m
Array length 37.5 m
Array width 5m
4. Data recording parameters
Recording system I/O systemII
Channels recorded 1440 data, 4 auxiliary
Record length 6s
Sampling period 4 ms
Tape recording format SEGD
Recording gain 48 db
Diversity staking Active
Stacking/correlation Correlation after vertical stacking
Fig. 7.27 Types of templates used in swath shooting. a Centre symmetric split. b Asymmetric split. c Double offend.
d Double within. e Centre multi shotline. f Centre of inclined shotline
More elaborate spread designs are based on shotlines. One of such designs is what is called
grouping the receiver points in blocks with the Checkerboard spread. An example represent
source points distributed regularly over parallel ing this type of spread is shown in Fig. 7.28.
7.3 3D Field Data Acquisition 159
(b)
(a)
Fig. 7.29 LoopSpread. a Survey points (source and receiver points) laid down around the irregular inaccessible area.
b Survey points laid down on the perimeter of a rectangle (or square) enclosing the survey area
160 7 3D Seismic Reflection Surveying
migration migration
aperture aperture
7.3 3D Field Data Acquisition 161
surrounding the target migrated area is required predened minimum to maximum offset values.
to insure full migration of the target area. The As a rule of thumb, an offset should be roughly
width of this strip (called the migration fringe) is equal to the depth of the target reflector. Thus, by
equal to the migration aperture which is dened dening the depths of target reflectors, it is possible
to be the horizontal distance a reflection event is to determine a range of offsets to suit both of the
moved by a migration process. The width of the targeted shallow reflectors and deep reflectors.
migration fringe is function of the dip measured For a given binsize, geophone group spacing
in the direction of the fringe width. The migra becomes known, and both total active channels per
tion fringe is shown in Fig. 7.30. receiverline and total number of receiverlines per
The survey area should be large enough to template can be decided, based on the maximum
include the fully migrated zone, migration fringe, offset required. Maximum offset (Xmax) and mini
tailfold zone, and the required surface mum offset (Xmin) can be calculated for the largest
coverage zone. For economic reasons, and when and smallest sourcetoreceiver points in the
the S/N ratio is sufciently high, the migration designed template (Fig. 7.31).
fringe may be extended in such a way as to include
(iii) The Swath Parameters
part of the tail zone. This step can result in
The main parameters of the swath are shooting
reduction of the total survey area to the area which
direction (survey orientation) and rollaside
is sufcient for achieving the targeted
overlap. Swath length and swath width are con
fullymigrated data volume.
trolled by the surveyarea boundaries and tem
Estimation of the fringe area is considered to
plate width respectively. Based on the
be an important issue in 3D surveys, since it is
foldcomputation formula, the fold is function
directly connected with the overall survey cost.
of both of the template movementoverlap and
The percentage costincrease due to addition of
swath roleaside overlap. Hence the overlap
migration fringe depends on the size of
(measured in terms of number of receiverstations
surveyarea as well as on the fringe width used.
for the template movement and in terms of
Cost saving, as we mentioned above, can be
receiverlines for swath movement) can be com
effected by allowing the fringe to be extended
puted once the fold is dened (Fig. 7.32).
into part of the tailfold zone. According to
The other important parameter related to the
(Sheriff and Geldart 1995, p. 452), the fringe area
swath is the shooting direction, or survey orien
should be equal to the migration aperture plus the
tation. When there is a distinct diptrend in the
radius of the rst Fresnel zone.
area, receiver lines are oriented in the direction of
(ii) Template Parameters dominant dip. In practice, this constraint is
The template is made up of a number of relaxed for two reasons: rst, dip value and dip
receiverlines, shotlines, number of active chan direction are generally not constant for the same
nels per receiverline, and number of sourcepoints area, and second, the dip moveout effect can be
per shot line. A template must be so designed that it taken care of, by dip moveout (DMO) correction
will provide a range of offsets covering a or by prestack migration in the processing stage.
162 7 3D Seismic Reflection Surveying
(iv) Bin Dimensions and Bin Attributes formula (Liner 2004, p. 272; Sheriff and Geldart
By denition, the bin is a subsurface rectangular 1995, p. 452):
(or square) area having length and width
Dx v=4f sinh
dimensions. These dimensions are equal to half
of the corresponding surface receiverstation
where (v = f k) is velocity, (f) is the highest
spacing. In processing, seismic traces are sorted
frequencycomponent, and (h) is the angle of dip
into bin gathers. This means that the bin will
component in the xdirection.
contain a number of traces (equal to the fold of
Derivation of the formula can be done with
coverage) of different offsets, and different az
the help of Fig. 7.33.
imuths. The centre of the bin marks the CMP at
This formula expresses the relation between
which the stacktrace of the bingather is located.
the CMP trace spacing (x) and the three vari
Thus, the bin area (normally referred to as the bin
ables (v, f, h). It has shown that (x), which is
size) should be small enough to achieve appro
representing the bin dimension in the xdirection,
priate resolution but not too small to create
is directly proportional to velocity (v) and
empty or lowfold bins. An optimum bin size
inversely proportional to wave frequency (f) and
must, therefore, be determined.
reflector dip (h).
The theoretical basis used for computing the
Typical bin dimensions used in land 3D surveys
optimum bin size is the same as that used for
are (12.525) m in the inline (subline) direction
avoiding spatial aliasing in case of 2D common
and (2550) m in the crossline direction. These
midpoint (CMP) spacing computations. Thus, in
ranges are corresponding to (2550) m and
order to avoid spatial aliasing, at least two sur
(50100) m of surface sampling intervals (receiver
face samples are required to be present for each
spacing). However, when square and small enough
apparent wavelength. For a given reflector dip,
bins are used, no spatial aliasing occurs. Small bins
the apparent wavelength should be as small as
mean ne subsurface sampling which is necessary
possible in comparison with the CMP spacing.
for increasing structural resolving power.c
From the geometry of the normal incidence
reflection from a dipping reflector and using Bin Attributes
(x = k/2 and t = s/2), a quantitative evalua The group of traces included in a bin gather, are
tion is obtained for the maximum spatial sam traces of different offsets recorded by receivers
pling interval (x), expressed by the following positioned at different azimuths. The number of
7.3 3D Field Data Acquisition 163
wavefront at
time, t
Sin =(t v/2) / x dipping
t = /2, = 1/f reflector
hence, x = v/4fsin
traces in the bingather represents the fold of Nyquist frequency is 250 Hz. which is expected
coverage for that bin. The parameters (offset, to be far higher than the highest frequency of
azimuth, and fold) form the bin attributes. recorded reflection signals. For economic rea
In the design process, it is aimed at sons, the recorded data is often resampled to
attributevalues that cover as wide coveragerange 4 ms before being inputted to processing. To
as possible, considering the economic constraints. make sure that no aliasing occurs, an antialias
Bin attributes are normally displayed in a special lter is applied in the acquisition recording stage
diagram (called spider diagram) in which offsets prior to AD conversion and it is also applied
are drawn as straight lines radiating from the bin prior to any resampling process carried out
centre. The line length is proportional to offset during processing work.
value and line bearing represents the azimuth. The sampling period, together with the bin
Number of lines is equal to the fold of coverage dimensions, set the three dimensions of the res
(Fig. 7.34). olution cell (the voxel) which has the general
In addition to the desired large coverageranges form of a parallelepiped or a cube. Its volume is
of attributes, the bin should be small and square to dened by the two horizontal dimensions (x
insure equallyne sampling in both of the and y), and the vertical dimension (vt/2),
subline (inline) and crosslinedirections. where (x and y) are equal to the bin dimen
sions in the subline and crossline directions,
(v) Sampling Period and Record Length
(t) is the sampling period, and (v) the seismic
The sampling period normally used in the seis
velocity (Fig. 7.35).
mic recording is 2 ms. With this value, the
XL1 XL2
(a) (b) (c)
SL2
y
x SL1
t
m
apex
circular m
wavefront
diffraction
s hyperbola
7.4 Processing of 3D Data 165
BING.
3. PARAMETER OPTIMIZATION
 Velocity analysis
 Residualstatic analysis
 Deconvolution analysis
 Filter analysis
4. POSTSTACK PROCESSING
 Zerophase deconvolution
 S/N Enhancement (fxy deconvolution)
 Bandpass filtering
 Equalization
 QC displaying
STKSEGY
166 7 3D Seismic Reflection Surveying
Fig. 7.38 The 3D data volume obtained from processing of a 3D dataset of an area of (10 by 20) km2
Fig. 7.39 Vertical sections (inline and crossline) obtained from a 3D data volume
168 7 3D Seismic Reflection Surveying
A seismic eld is created as a result of a sudden nding a way to apply statistical communication
mechanical impact that occurs at a point inside a theory to the seismic problem (Enders Robinson
medium. The created packet of energy (called the 1983, p. 221).
seismic energy pulse) spreads out from the The GAG research activities have resulted in
source zone in all directions. It travels through furnishing a way for applying the principles of
the medium in such a way that energy moves the communication theory in the analysis of
through the particles of the medium without seismic wavelets. In essence, the seismic wavelet
creating any permanent changes to the medium. is considered as a propagating signal which can
Propagation of the generated seismic energy in be processed using principles of the communi
this way is expressible in terms of a mathematical cation theory in the same way as processing of
equation, the (wavemotion equation). the electromagnetic signal.
In a seismic reflection exploration that uses The GAG scientic ndings in the late 1950s
impulsive energysources, the generated seismic have set the cornerstone of modern digital pro
pulse is generally made up of an oscillating cessing of the seismic signal (the seismic
energy level that diminishes with time and with wavelet).
travel distance, forming what is normally called
the (seismic wavelet). This travelling pulse (the
wavelet) can be recorded by placing a seismic 8.1 Definition of the Seismic Signal
detector (seismometer) in, or on the surface of,
the medium. The seismic wavelet detected near In general, the travelling seismic wavelet (de
the source location is normally referred to as the tected directly, or after being reflected or refrac
source wavelet (or source signature), and that ted) is considered to be a travelling signal that
detected after being reflected from an interface, it bears useful information. The seismic reflection
is called the reflection wavelet. signal, normally representing the particle
In 1952, Massachusetts Institute of Technol vibration velocity, is a function of space and
ogy, MIT (USA) set up and sponsored a research time. The value of the function at any time is
group, called the Geophysical Analysis Group dened to be the amplitude (or energy) of the
(GAG) which was in 1953 taken over by a signal. Usually, the seismic signal is presented as
consortium of oil and geophysical companies. a mathematical function that expresses amplitude
This group of scientists had the responsibility of variation with time (Fig. 8.1).
Table 8.1 Changing factors of the seismic reflection signal during its sourceto receiver travelpath
Changes in source zone Changes in raypath zone Changes in receiver zone
Source type of energy source Inelastic attenuation Receiver detection response
Source coupling Geometrical spreading Detector coupling
Nearsource geology Reflection coefcient Nearreceiver geology
Sourcegenerated noise Wave conversion Noises and interferences
Noises and interferences
8.1 Definition of the Seismic Signal 171
f(t)
F()
t
1
g(t)
G()
t
2
t 1 2
Fig. 8.2 Concept of the frequency spectrum. The monofrequency functions, f(t) and g(t) have one line spectrum each.
Their sum, f(t) + g(t), have twoline spectrum
A0 time (t)
A2
A( )
a0 a1 a2 a3 a4 a5
A3
1 2 3 4 5
A4
phase spectrum
()
A5
Fig. 8.3 Pictorial representation of spectrum analysis. The signal (at top of gure) is Fouriertransformed into its
frequency components shown as amplitude and phase spectra, A(x) and (x)
This simplied presentation represents a syn present in the analyzed signal. Two spectral parts
thesis process by summing two sine functions. are necessary to completely dene the complete
The reverse of this process is the spectrum anal signal characteristics. These are the frequency
ysis which involves (using Fourier transform) spectrum and the phase spectrum. The two con
determination of all of the frequency components cepts are schematically shown in Fig. 8.3.
of a given signal. The complete analysis is to get The frequency range of the seismic reflection
amplitudes and phase shifts of all frequencies signal is typically (2060) Hz.
172 8 The Seismic Reflection Signal
8.1.3 Comparison with the Radar conditions. The manner and severity of these
Signal changes depend largely on:
Geometrical shape of the traversed layers
Radar signal and seismic signal are similar, in that (structural changes)
they both provide the sourcetoreceiver distance Physical properties of the layers (stratigraphic
of a faraway object, using the wavereflection changes)
phenomenon. The source signal used in the case of Fluid content like water, oil, and gas (reser
radar is typically a short pulse (of a wide voir conditions)
frequencyband), a chirp signal (linearlychanging The seismic wavelet is therefore considered as
frequency signal), or a frequencymodulated a messagebearing record of the changes which
(FM) signal. The radar function is based on two are coded in the wavelet spectrumparameters
main types of measurements; the traveltime and (amplitude, frequency, and phase). In the
Dopplereffect measurements. The Doppler Effect timedomain language we say that we have
(variation of frequency with target motion) is waveformdeformation whereas in the frequency
made use of, in velocity measurement. The com domain this is described as spectraldeformation.
mon feature between radar operation and seismic Compared with the source function, and due
reflection technique is that both of them use the to these affecting factors, the detected reflection
reflection traveltime measurement. Naturally, signal becomes lower amplitude, lower fre
Doppler Effect has no application in seismic quency, extended in shape, noisy, and
reflection operation. timeshifted with respect to its initial origin time.
The main differences between the seismic Mathematically, these changes can be classied
signal and the electromagnetic signal used by into the following four types of processes:
radar are summarized in Table 8.2. superposition, scaling, ltering, and time
This table shows that the parameters of the shifting.
seismic signal are dispersed over a wide range of (i) Superposition
values, compared with those of the radar signal in Interaction of the reflection seismic signal
which the parameter has only one value. This is with noises and other interfering events
because the radar signal is travelling through a (such as coherent and incoherent noise,
medium which is far more homogeneous than refractions and diffractions) is a simple
that traversed by the seismic signal. superposition process. The received signal
is resulting from additions of all of these
interferences, bringing about distortions to
8.1.4 The Seismic ReflectionSignal the wavelet geometrical shape.
Changes (ii) Scaling
This is an arithmetic multiplication pro
The seismic reflection wavelet is, in fact, the cess, where the reflection wavelet is mul
source pulse after being modied and distorted as tiplied by a constant scaling factor, similar
a result of the interactions with the travelpath to the effect of gainapplication. The effect
8.1 Definition of the Seismic Signal 173
Fig. 8.4 Summary of the  high amplitude (i) superposition  lower amplitude
physical changes of the  high frequency content (ii) scaling  lower frequency content
reflection signal as it travels  sharp shape (iii) filtering  extended
from the source point (S) to  with no noise (iv) time shifting, to  noisy
the detector point (R), via  Onsettime, t=0
 time Shifted, t=t0
the reflection point (RP) S R
source impulse reflection wavelet
V1
s(t) r(t)
t V2 t
0 0 t0
V3
V4
RP
does not deform the signal but only detector offset from the source.
change its size. Examples of this type of Timeshifting becomes a deformation
effect are the effect of the geometrical factor when reflectors are so close to each
spreading and the reflection coefcient other that waveletarrivals overlap causing
which are independent of frequency. This distortions to the wavelets and loss of the
means that they incur no wavelet shape resolution of the reflection events. Phase
deformation. A scaling effect can, under shifts are the time shifts of the individual
certain conditions, reverse the wavelet frequency components which, in general,
polarity. This happens when the scalar is take place in frequencyltering actions.
of negative value, as in negative reflection These changes are summarized in Fig. 8.4.
coefcient.
(iii) Filtering
Frequency ltering can be considered as a 8.2 The Seismic Trace
frequency selective scaling. A typical
example of this type of effect is the earth A seismic pulse incident at an interface is partly
highcut lter, which is operating on the reflected and partly transmitted into the following
seismic wavelet attenuation throughout its layer. In a multilayer medium this is repeated at
travelpath. Any effect that is each interface present in the way of the advanc
frequencydependent (as the earth lter ing wave. A detector placed on the surface
ing) incurs shapedeformation to the receives sequentially, the reflected wavelets, at
propagating seismic wavelet. The defor timeintervals depending on the depths of the
mation severity depends on the type of reflectors. The source pulse reaches the detector
lithology and fluidcontent of the geolog after being affected by the various types of
ical formations traversed by the seismic modications (as discussed above) including the
signal. Being an analoguetype, the earth process of reflection in which amplitude and
ltering effect introduces phase shift to the polarity are governed by the reflection coefcient
frequency components of the seismic of the reflector. From each reflector a wavelet
signal. representing the source wavelet, scaled by the
(iv) TimeShifting reflection coefcient, will arrive at the detection
The source impulse takes time to reach the point. Thus, a series of such wavelets, each of
detector. This is the total reflection which is shifted in time with respect to the pre
traveltime which is depending upon the ceding wavelet, are superimposed on each other
depth of the reflector as well as on the to form a record which is the (seismic trace).
174 8 The Seismic Reflection Signal
Reflector1
1
Reflector2
2
Reflector3
3
Reflector4
4
Reflector5
5
Fig. 8.5 Sketch of the convolutional model of the seismic trace (Alsadi 1980, p. 192)
Seismic trace
Source function
Earth filter
Reflectivity
Instrument. Resp.
Geom. Spreading.
Noise
Fig. 8.7 The seismic
reflection trace. a Pictorial (a) (b)
representation. b Actual amplitude amplitude
seismic trace from a
0
dynamite sourceenergy
1
reflection time
2
in
3
seconds
8.2.2 The Synthetic Seismogram The earth layering model is normally obtained
from a drilled explorationwell. The well logging
A direct application of the concept of the con data furnish both of the velocity (from the sonic
volutional model is in constructing an articial log) and density (from density log). From these
seismic trace, commonly known as the (synthetic two logs, the acoustic impedance of each layer is
seismogram). The basic requirements for this calculated, and hence the reflection coefcients
process are the source wavelet and the reflectivity of the layer boundaries are determined. This is
series, which in turn, require the distribution with the reflectivity series which will be in the form of
depth of the layers acoustic impedances. values, ranging between (1) and (+1),
176 8 The Seismic Reflection Signal
(x) (*)
representing the reflection coefcients of all of with the real geological section. It is normally
the interfaces (layerboundaries) penetrated by displayed (using several duplicated synthetic
the well. Convolution of the source wavelet traces) superimposed on the seismic stack section
(assumed or obtained from a nearby shotrecord) at the welllocation. Main purpose is to attach the
with the reflectivity series, gives the synthetic stratigraphic identities for the reflection events
seismogram. This computation method expresses diagnosed on the seismic stack section (Fig. 8.9).
the convolutional model according to which the
real seismic trace is assumed to have taken place
in real time seismic reflection recording. 8.2.3 The Digital Seismic Trace
In display of the computed synthetic seismo
gram, the time scale must be doubled in order to The seismic trace received at the detection point
match the seismic section in which the time is a is represented as a continuous amplitude varia
twoway vertical time. A schematic representa tion with recording time. At this point it is an
tion of the process is shown by Fig. 8.8. analogue function expressing the seismic ampli
Interpreters are usually provided with syn tude as function of time. The recordingsystem
thetic seismograms both with and without multi transforms the amplitudeversustime function
ples to help in sorting out the multiplereflections, from its analogue form to digital form, and gets it
if any. stored on the magnetic tape as a digital function
The synthetic seismogram is now considered (sequence of samplevalues at regular time
as an indispensable tool in the hand of the intervals, t), where the sampling period (t) is
interpreters to tie the seismic reflection images normally chosen to be 2 ms (Fig. 8.10).
8.2 The Seismic Trace 177
TWT
With the digital mode, the seismic trace stor 8.3.1 Definition of the Seismic
age and processing become feasible to carry out Wavelet
by electronic digital computation systems.
A wavelet is dened as a transient signal with a
denite time origin. It is characterized by two
8.3 The Wavelet Concept properties; it has a denite onset time and it has a
nite energy (Robinson 1983, p. 128). These two
The seismic pulse created by a mechanical shock characters implies that the wavelet is onesided
(seismic source) is a wavelet that is transmitted entity (wavelet values are zero before onset time),
through the earth medium and detected by the and it is transient with its energy diminishing after
receiver geophonegroup and then recorded by onset time. Ricker wavelet (named after Norman
the recording system. At the time it is recorded, Ricker 18971980) is a special type of a
the wavelet is no longer a short pulse (source theoreticallycomputed zerophase wavelet com
impulse) as it was created at the source location, monly used in seismic modeling studies. In par
but an extended wavelet madeup of few cycles, ticular, Ricker wavelet is used in the seismogram
weakened pulse having energy diminishing to synthesis, by convolving it with the reflectivity
zero level within a short time (tens of millisec function, to produce the nowaccepted convolu
onds) (Fig. 8.11). tional seismic trace model (Robinson 1983, p. 224).
Functions of geophysical signals can be repre The process of converting an analogue function
sented either as continuous function (analogue into digital form (AD conversion) is normally
function) in which the function is dened at all of referred to as (digitization or sampling) process.
the points along the abscissa or dened at discrete It involves twosteps; sample denition followed
points, normally regularly spaced values (digital by quantization and coding in which the sample
function). Considering timedomain functions, value is determined and converted into digital
the analogue function is normally quoted as s(t), number. The output of the sampling process is
and the digital function as s(n t), where t is the then stored on a suitable digital storage device or
sample spacing along the timeaxis, normally entered in a certain digital data processing
called the sampling period. The two forms of operation. In the language of mathematics, the
functions are shown in Fig. 8.13. AD process is represented as follows:
t t
0 0
analogue Function, s(t) digital Function, s(n. t)
s(n. t) = 5 3 1 5 3 2 0 7 12
n = 3 2 1 0 1 2 3 4 5
s(n. t) = { 2 5 3 1 3 0 2 5 7 }
s(n. t) = { 2 5 3 1 3 0 2 5 7 }
s(n. t) = { 2 5 3 1 3 0 2 5 7 }
3 2 1 0 1 2 3 4 5 6 7 8
8.5 The Sampling Theorem that sampling which produces at least two sam
and Aliasing ples per period of the periodcomponent having
the shortest period in that function. The concept
Digitization of a signal implies reading the signal
is shown in Fig. 8.15.
at isolated points which means that the infor
In the frequencydomain language the theo
mation of the signal in between these points is
rem statement can be rephrased as, it is possible
permanently lost. The amount of information lost
to completely recover the original analogue
is naturally greater as the sampling period is
function from its digital form if the sampling
larger. In the sampling process, the shortest the
frequency (fs) is greater than twice of the highest
sampling period, the better is the recovery of the
(maximum) frequency (fm) present in the original
details of the function. In fact, the analogue
182 8 The Seismic Reflection Signal
t t
t = /2
fs 2 fm
8.5.2 The Aliasing Phenomenon
There is another term which is intensively There are three categories of sampling; ne
used in connection with signals sampling, called sampling, critical sampling, and coarse sampling,
Folding frequency or Nyquist Frequency, (fN), depending on the magnitude of the Nyquist fre
which is dened to be half the sampling fre quency (fN) relative to the cutoff frequency
quency, that is: (maximum frequency component, fm) present in
the original analogue signal. The categories are:
f N f S =2 1=2Dt (i) Fine Sampling (fN > fm)
In this case high sampling frequency is
With the introduction of the Nyquist Fre
applied where (fN > fm) complying with
quency (also called folding frequency), the con
the sampling theorem and complete
dition in the sampling theorem can be restated as
recovery of the original analogue signal
(For complete recovery of the original analogue
can be obtained, with no distortions.
function from its digital form, the Nyquist fre
(ii) Critical Sampling (fN = fm)
quency must be greater than the highest fre
As in the nesampling case, no distor
quency present in the original signal).
tions occur in this case. It is critical, in the
The important conclusion extracted from the
sense that the applied sampling frequency
Sampling Theorem is that, Nyquist frequency,
sets the marginal limit of the Nyquist fre
used in sampling of an analogue signal, must be
quency below which distortion (frequency
equal or greater than the highest frequency pre
aliasing) shall take place.
sent in that signal, that is:
8.5 The Sampling Theorem and Aliasing 183
fm
fA
Fig. 8.16 The aliasing phenomenon. Input frequency (fin) = 100 Hz, Sampling frequency = 120 Hz, Nyquist
Frequency = 60 Hz, Aliasing frequency (fA) = 20 Hz
(iii) Coarse Sampling (fN < fm) frequency (fN f). This means that if the input
With coarse sampling (or undersampling, to the sampling process is (fN + f), the output
as it is sometimes called), those compo frequency will be (fN f). The two frequencies
nents which are of greater frequencies than (fN + f) and (fN f) are called aliases of each
Nyquist frequency will appear as lower other (Sheriff 1973, p. 4). Thus, for example, a
frequencies in the sampling output. This 100Hz signal is sampled at sampling frequency
phenomenon, which occurs only when of 120 Hz (fN = 60 Hz), then the output will be:
coarse sampling is applied, that is when 60 (100 60) = 20 Hz, which is the aliased
(fN < fm), is called (aliasing). An example frequency (Fig. 8.16).
of an aliasing case is shown in Fig. 8.16. The general rule governing any sampler
inputoutput relationship is that any frequency
component higher than the Nyquist frequency
8.5.3 The Aliasing Frequency (fN), present in the signal prior to sampling, is
Computation outputted as alias frequency. The sampling fre
quency (fS), or the Nyquist frequency fN (=fS/2),
In any sampling process, the frequency compo is the deciding factor for the possibility of
nents in the input analogue function which are occurrence of aliasing. Cases for the ne, critical,
higher than Nyquist frequency will appear at the and coarse sampling of an input frequency of
sampler output as false components (aliased (125 Hz) are presented in Table 8.3.
components) of frequencies lower than input In Table 8.3 one can notice that, in the rst
frequency. An input frequency component (fin) two cases (belonging to the case of
higher than Nyquist frequency by (f) will nesampling), the Nyquist frequency is higher
appear to the sampling system, as the lower than the input frequency and hence no aliasing is
Table 8.3 Cases for ne, critical, and coarse sampling of an input frequency of (125 Hz)
Input fm (Hz) Sampling freq. (Period) Nyquist freq. (Hz) Output freq. (Hz) Output status
1 125 500 Hz (2 ms) 250 125 no aliasing
2 125 333 Hz (3 ms) 166.7 125 no aliasing
3 125 250 Hz (4 ms) 125 125 no aliasing
4 125 200 Hz (5 ms) 100 75 with aliasing
5 125 166.6 Hz (6 ms) 83.3 41.6 with aliasing
6 125 142.8 Hz (7 ms) 71.4 17.8 with aliasing
7 125 125 Hz (8 ms) 62.5 0.0 with aliasing
184 8 The Seismic Reflection Signal
0
input frequency, fin
Fig. 8.17 Inputoutput relationship, in a sampling process. The zone up to the Nyquist frequency (fN) is the
principalaliases zone (PA) in which no aliasing occurs
taking place. In the third case (the applied with no aliasing effect. An input fre
criticalsampling case), Nyquist frequency is quency, lower than the Nyquist frequency, will
equal to the input frequency and again no alias be outputted with no aliasing effect.
ing occurs. In cases (4, 5, and 6) Nyquist fre The periodic linear relation of the sampler
quency is less than the input frequency, hence inputoutput function, shown in Fig. 8.17, can be
aliasing takes place. In the last case (case 7) the used in deriving a general formula that can be
difference (Df) has reached its maximum value used to calculate the output aliased frequency,
which is equal to the Nyquist value giving an given the input frequency and the sampling (or
output of zerofrequency. Nyquist) frequency. With reference to this gure
To give a quantitative measure for the output it can be shown that the output aliased frequency
frequency, use is made of the periodic linear (fout) is related to the input frequency (fin) by the
relation connecting the input frequency (fin) and following relationship:
output frequency (fout), as shown in Kanasewich The integer (k) is calculated from floor
(1973, p. 89). Because of the feature that the truncation of (fin/fN), which is symbolically
spectrumpart above the Nyquist frequency can represented as shown above, that is,
be folded back about the Nyquist frequency, k bf in =f N c
aliasing is sometimes called (folding) and the For application, let us nd the output aliased
aliased frequency, called folding frequency. The frequency for the input frequency, (fin = 782
(fin fout) relation is reproduced in Fig. 8.17. Hz), sampled at 2ms sampling period, i.e. at
This gure is helpful in determining the sampling frequeny (fS = 1/0.002 = 500 Hz). The
inputoutput relation for any sampling process, Nyquist frequency, fN = 500/2 = 250 Hz, and
once the sampling frequency (fS), and hence, the the integer (k) is obtained from truncation of
Nyquist frequency (fN) is known. In the zone (0 (fin/fN = 782/250 = 3), hence, fout = 782
fN), the output frequency (fout) is the same as [3 + {1 (1)3} 250 = 218 Hz. Similarly,
the input frequency. This region of the spectrum, it is possible to calculate the aliased frequency of
called the principal aliases section, marks the any other input sampled signal having frequency
limits of the sampling frequencies which can be exceeding the Nyquist frequency. As examples
8.5 The Sampling Theorem and Aliasing 185
for this method of calculation, let us work out the 8.5.5 Aliasing in the Frequency
aliased frequencies for the input frequencies Domain
(fin = 457, 557, 957, 1957) Hz, each of which is
sampled at 2 ms sampling period. The output The analogue function can be looked upon as a
aliased frequencies for any one of these fre digital function sampled at zerosampling period,
quencies will be (fout = 43) Hz. which will have an innitelylong spectral repe
tition cycle. This is equivalent to saying that its
spectrum is of nonrepetitive nature in case of an
8.5.4 Effect of Sampling on Signal analogue signal. When the analogue is converted
Spectrum into digital form the amplitude spectrum is
repeated over the frequency axis at
The direct effect of the sampling process on the repetitionspacing equal to the sampling fre
frequency spectrum of a signal is generation of quency (fS). The aliasing problem occurs when
repeated spectrumreplicas of the original ana the cutoff frequency (fm) of the analogue signal
logue signal. Theoretical computations have exceeds that of the Nyquist frequency (fN), that is
proved that Fourier transform of a digital function when (fm > fN). In this case, the timedomain
is itself periodic with repetition interval equal to aliasing becomes spectrum overlapping in the
the sampling frequency used in the sampling frequency domain with the consequence of
process (Kansewich 1973; Kulhanek 1976). Thus, spectrum distortion due to the resulting
after sampling of an analogue signal, its spectrum spectruminterference. Aliasing, in the frequency
gets repeated over the frequency axis at regular domain, is, for this reason, called spectrum
spacing which is equal to the sampling frequency contamination (Bth 1974, pp. 148151).
(1/t), where (t) is the sampling period. The important point here is that when the
The lower the sampling frequency the smaller input analogue signal contains frequency com
the spacing becomes. With toolow sampling ponents higher than the Nyquist frequency
frequency (too large t), the spectrum spacing (fm > fN), these high frequencies will fold over
gets smaller and when Nyquist frequency and add to those frequencies existing below the
becomes less than the cutoff frequency of the Nyquist frequency. The outcome of the process
signal (fN < fm), the repeated spectra overlap and in this case is spectrum distortion in the overlap
spectrum distortion occurs. zone as shown in Fig. 8.19.
The spectrum behavior caused as a result of
digitization is shown in Fig. 8.18.
1/t
digital function repeated spectrum
186 8 The Seismic Reflection Signal
freq. freq.
fN fN
8.5.6 The Remedy for the Aliasing than the Nyquist value, and hence, aliasing is
Effect avoided.
In summary, in order to avoid aliasing, there
To avoid aliasing and its distortion consequences are two alternative methods:
a procedure should be taken such that the con
(i) Applying higher sampling frequency (small
dition (fN > fm) is restored. To achieve this state,
sampling period) up to the value with which
one can either delete those frequencies which are
the Nyquist frequency becomes greater than
greater than Nyquist frequency prior to the dig
the cutoff frequency of the input analogue
itization process, or increase the sampling fre
signal.
quency in such a way as to get a Nyquist
(ii) Applying a high cut lter (the antialias
frequency of greater value than the cutoff fre
lter) to remove those high frequencies
quency of the original analogue signal.
(from the original analogue signal) which
In practice this is done by application of a
are higher than the Nyquist frequency
suitable highcut lter (usually referred to as
before starting the sampling process. The
antialias lter), to the input seismic data prior
two methods are explained in (Fig. 8.20).
to the sampling process. In effect, the applica
tion of the antialias lter makes the cutoff In seismic reflection datarecording, the sam
frequency of the input analogue signal, lower pling period is normally set at 2 ms value. This
makes the Nyquist frequency to be 250 Hz. This vertical resolution and horizontal resolution (ex
means that with this Nyquist value, no aliasing plained herebelow).
effect is expected since the maximum frequency
(signal cutoff frequency) is expected to be far
below (250 Hz). However, all recording systems 8.6.1 Vertical Resolution of Seismic
are equipped with the antialiasing lter that can Signals
be applied when it is required. The antialiasing
lter is normally applied whenever resampling Vertical resolution of seismic reflection events, is
of the data is carried out. Very often, in seismic dened as the minimum vertical distance
data processing, the recorded digital data is between two interfaces that give two distict
resampled from 2 to 4 ms sampling period for reflection events on a seismic section. It is basi
economic motives. cally governed by the wavelength of the seismic
signal. The shorter the wavelength (i.e. the higher
the frequency) the greater the vertical resolution.
8.6 Signal Resolution In addition to the frequency factor, depth and
and Resolution Power reflector spacing have signicant effects on the
resolution. The vertical resolution is governed by
Resolution is dened as the ability of distin the ratio of the depth separationdistance of the
guishing individual objects gathered to gather in reflectors (z) to the wavelength (k) of the
one group, or the details of shape changes of an incident seismic signal. The lowest limit (reso
irregularly shaped object. In the eld of seismic lution limit) of this ratio; (z/ k) is found to be
exploration, the seismic resolution is the ability (1/4) (Sheriff and Geldart 1995, p. 174). This
of recognizing two adjacent seismic events as means that the reflector separation must be more
distinct two events and not as one blurred event. than quarter of the wavelength (z > k/4). This
Resolution power can be measured by the also means that the image separation (events
minimum separation distance between two seis separation measured on a seismic stack section)
mic events that can be resolved as two distinct should be more than half a period (t > s/2) in
features on the seismic section. Obviously the order to be distinctly resolved (Fig. 8.21).
sharper the reflection wavelet, and higher Adoption of the (k/4) criterion for the resolu
signaltonoise ratio, the better the resolution tion limit, implies that the two events reflected
power will be. The term (resolution power) is from two neighboring interfaces are separated by a
used to imply ability of detecting and bringing to half cycle, which means that depth separation (z)
vision a certain seismic event between two neighboring interfaces greater than
In seismic work, we are concerned with two (k/4) will lead to minimum destructive
types of resolutions of seismic reflection events:
z = v. t/2 = v. (/2)/2 = /4
188 8 The Seismic Reflection Signal
interference between the reflected waves from the earth highcut lter with depth, both are
two interfaces, causing resolution deterioration. leading to increase of the wavelength and
hence decreases the resolution power.
(i) Factors Affecting Resolution
Application of highcut lters (as
In general the higher the frequency content
application of the antialiasing lter),
of the seismic trace the better is the reso
which lead to attenuation of the high
lution power. Well logs (wireline logs)
frequencies of the signal and hence
have greater resolution power than seismic
result in lowering the resolution power.
traces since well logs are generated by high
(ii) The problem of Thin Beds
frequency sources. These logs can resolve
A special case, related to the subject of
beds on centimetermeter scale while seis
resolution which brought appreciable
mic reflection records cannot resolve
attention by geophysicists, is the problem
somuch detailed variations. Reflection
of resolving thin beds. Two reflectors
survey data can resolve reflectors at
spaced by less than quarter of a wave
depthseparation of about 10 m at its best.
length, have reflection responses depend
The main factors affecting resolution are
ing on the layering model. A layer is
reflector spacing, reflector depth, and
regarded as a thin layer when its thickness
reflection signal frequency. Closely spaced
is less than a quarter of the dominant
reflectors cause interferences of reflected
wavelength (Sheriff 2002, p. 353).
waves which lead to loss of resolution. The
Consider a thin bed of thickness of (k/4)
resolutionpower is generally decreasing
and of velocity (V2), sandwiched between
with depth for the following reasons:
two layers of velocities (V1 and V3) where
Earth lter which is cutting high fre
(V1 = V3 < V2), the wave reflected from
quencies, that is cutting short wave
its top and that from its base will interfere
lengths. Thus for depth of 800 m, say,
constructively producing a highamplitude
velocity of 1000 m/s, and frequency of
reflection, forming what is normally
100 Hz, the wavelength will be 10 m
referred to as the thinbed effect or tuning
and the resolution becomes only 2.5 m.
effect. If, on the other hand, the thinbed of
However, when depth is 3000 m,
velocity (V2) found between two layers of
velocity of 4000 m/s and frequency of
velocities (V1 and V3) where (V1 < V2 <
25 Hz, the wavelength will be 160 m
V3), destructive interference will result
and the resolution becomes 40 m. In
(Sheriff and Geldart 1995, p. 174). These
general resolution gets less (poorer
two models are shown in Fig. 8.22.
resolution) with increasing depth due
Vertical resolution is always improved
to the effect of the earth highcut lter.
with higher seismic frequencies. But, due
Increase of velocity due to compaction
to the earth ltering effect, frequencies get
and decrease of frequency due to the
lower with increase of reflector depth.
(a) (b)
V1 = V3 < V2 V1 < V2 < V3
V1 V1
V2 V2
V3 V3
Fig. 8.22 Two models of a thin bed, having interval velocity (V2). Model A (V1 = V3 < V2) and model B
(V1 < V2 < V3)
8.6 Signal Resolution and Resolution Power 189
Table 8.4 Estimates of the minimum depth interval resolved, corresponding to thre values of sampling period
Sampling period (ms) Highest frequency (Hz) Wavelength (k) (m) Min. depthInterval (z = k/4) (m)
1 500 5 1.25
2 250 10 2.50
4 125 20 5.00
This table shows that it is always possible to Decimal 105 104 103 102 101 100
devise a new numbering system, but the most system
suitable for us is the decimal system and most Binary system 25 24 23 22 21 20
suitable for computers (computations and stor Octal system 85 84 83 82 81 80
age) is the binary system. Hexadecimal 16 5
16 4
16 3
16 2
16 1
160
system
where R is the radix, and the letters (A, B, C, ) 8.7.2.1 Conversion to Decimal System
are the digits of the number. Conversion is done by applying the mathematical
The following table shows the concept sequence that denes the numbering system
applied to the four most commonly applied concerned. Examples of the conversion process
numbering systems: are given in the following examples:
192 8 The Seismic Reflection Signal
47 D 101111 B 57 Oct
and
Examples:
[ 001 001 010 001 . 100 101 ]B = [ 1121.45 ]O
It is to be noted here that there is yet another lesscommon numbering system called the
binarycoded decimal (BCD) numbering system. In this system, each of the decimal digits is sub
stituted by the corresponding group of the binary digits. Each of these groups consists of 4 binary
digits, as in the following example:
There is another method to do the subtraction To clarify the process let us take the example of
process, and that is by use of what is called subtraction of the number [+53]D from the number
(Twos Compliment method). The subtraction [+67]D, assuming that the subtraction process is
process starts with converting the number to be done by an 8bit computer. This is done as follows:
8.7 The Common Numbering Systems 195
[ 67 ]D = [0 1 0 0 0 0 1 1 ]B
[ 53 ]D = [0 0 1 1 0 1 0 1]B
For an 8bit computer, the digit (1) appearing the decimal numbering (1, 2, 3, ,). Note that after
after the 8th digit will be outside the computer the digit (9) in the decimal system comes (10) and
binaryword limit and so will be neglected and the after the digit (F) in the hexadecimal system
result will be the 8digit word (1110) which is [14]D. comes (10), and after the digit (7) in the octal
system comes (10), and after the digit (1) in the
binary system comes (10). This means that, after
8.7.3 Counting in the Different the complete set of digits of each system comes
Numbering Systems (10), then counting continues. The following table
shows this method applied for the hexadecimal,
Here is a table of different numbering systems decimal, octal, and binary systems.
(hexadecimal, decimal, octal, and binary). Each
column contains the numbering corresponding to
0 10 20 30 . 0 10 20 30 . 0 10 20 30 . 0 10 100 110 .
1 11 21 31 . 1 11 21 31 . 1 11 21 31 . 1 11 101 111 .
2 12 22 32 . 2 12 22 32 . 2 12 22 32 . .
3 13 23 33 3 13 23 33 3 13 23 33
4 14 24 34 4 14 24 34 4 14 24 34
5 15 25 35 5 15 25 35 5 15 25 35
6 16 26 36 6 16 26 36 6 16 26 36
7 17 27 37 7 17 27 37 7 17 27 37
8 18 28 38 8 18 28 38
9 19 29 39 9 19 29 39
A 1A 2A 3A
B 1B 2B 3B
C 1C 2C 3C
D 1D 2D 3D
E 1E 2E 3E
F 1F 2F 3F
The Seismic Processing Tools
9
From mid 1950s till mid 1960s a jump in the In order to extract the useful message (geological
seismic data processing took place. Within this information) from these recorded data, these
interval, the analogue magnetictape recording seismic traces are subjected to a set of analysis
and processing, followed by the introduction of procedures, using certain processing tools. These
digital technique, were introduced. A third and tools are numerical analysis techniques applied
equally important factor that boosted the seismic on the input seismic traces. The most common
data processing is the adoption of concepts bor tools applied in seismic data processing are:
rowed from the communication theory which
Fourier Analysis
considered the seismic wavelet as a travelling
Correlation functions computations
signal similar to the electromagnetic signal.
Convolution (frequency ltering)
The input data to the processing system con
Deconvolution (inverse frequency ltering)
sists mainly of the digital seismic traces, recorded
Frequency and velocity ltering
by the eld recording system. The fundamental
Equalization (trace scaling)
principle on which processing is based upon, is
Sample Editing (timeshifting, signchanging,
that the seismic trace is considered to be as a
sample zeroing)
digital signal of value (amplitude) which is
function of the reflection travel time. In order to These processes form the basic processing
extract the useful information the seismic signal tools in the hand of the geophysicist, ready to be
(seismic reflection wavelet), a number of math used in processing of seismic reflection data.
ematical and statistical processes, are applied.
These processes are considered to be the tools 9.1.1 The Sine Function
employed to do the required analyses.
The building brick of any physically realizable
function is the sine function. According to Fourier
9.1 The Seismic Processing Tools theorem, a time or distancefunction, is made up
of a sum of innite number of sinefunctions. Due
The raw seismic data acquired in the eld, which to its important role in processing in general, and
form the input to the processing system, are in Fourier analysis in special, an introductory
normally digitally recorded traces on special treatment of the sine function shall be given
magnetic tapes. The reflection wavelets received herebelow, preceding the discussions of the
by a surfacepositioned detectionsystem (the processing tools listed above.
geophone group) are digitally recorded as func One way to generate a sine function is by using
tion of recording time forming the seismic trace. a geometric approach. Let us consider a unit circle
P1 P1 360
P
P2 P0 P0 P2
O R 90 180 270 360 450
P3 P3
of radius equal to unity, in which the radius is relation between the two unit systems; degrees
rotating with constant rotation speed about the and radians:
circle centre. At a given time, the radius (OP in
Fig. 9.1) makes an angle (h) with the xaxis, and 360 2p rad
its tip, point (P) has its projection on the xaxis
represented by point (R). By denition, the sine Given (p = 3.1415927), the relation between
function (sin h) has a value given by the length the two systems is that 1 rad = 57.2958, or
(PR). Now, as the radius rotates about the centre 1 = 0.017453 rad. The argument of sine func
point (O) the length, PR (=sin h) varies with the tions (h, in this example) is normally quoted in
angle (h). By drawing the value of the PRlength radians.
as function of (h), we obtain the curve for the sine
function (sin h), as it is shown in Fig. 9.1.
9.1.3 Parameters of the Sine
Function
9.1.2 The DegreesRadians
Relationship The general form of the sine function is that it is
function of angle (h). Other common forms of
Referring to Fig. 9.1, the point (P) is rotating at this function are: function of time s(t) and
constant speed and the generated curve is made function of distance s(x), thus:
up of repeated cycles corresponding to the
repeated rotation cycles of the point P around the sh a sinh
circle circumference. The generated (sin h curve)
repeats regularly every 360. The complete angle or,
swept by the rotating arm is from (h = 0) when P
st a sin2pf t t tp ; f t 1=s
is at location (P0) to (h = 360) when P is back at
location (P0) passing through all of the points or,
(P0, P1, P2 and P3).
It should be noted here that angles can be sx a sin2pf x x xp ; f x 1=k
expressed in radians rather than in degreeunits.
An angle (h) in radians is dened as the ratio where, (a) is amplitude, () is phase angle, (ft) is
between the subtending arc divided by the radius temporal frequency, (fx) is spatial frequency, (s)
of the circle in which the angle is central. When is period, (k) is wavelength, (tp) is phase shift in
the arc length is equal to the radius, the angle time units, and (tx) is phase shift in distance
value is 1 rad. Since circumference of a circle is units.
equal to 2p times its radius, the value of the total The maximum value of the sine function is
central angle is 2p. This leads to the important equal to the length of the rotating arm applied for
9.1 The Seismic Processing Tools 199
generating the sine function considered above. arm sweeps a complete central angle (360),
A sine function, represented by a sine curve which is (2p) when expressed in radian units.
(sometimes called sinusoid) is completely When frequency is expressed in terms of rate of
dened by three parameters: amplitude (a), fre change of the angle (dh/dt), measured in radians
quency (ft or fx) and phase (, tp, or xp). The per second, it is usually called, radian frequency.
amplitude is of positive value (+a) represents a Thus, for constant rate, we can write:
peak and negative value (a) represents a trough
of the given sinusoid. x h=t 2p=s 2pf
The most common form of the sine function is
the form a sin(2pft + tp) quoted as function of where, (x) is the radian frequency (rad/s) and
time (t). The frequency, (f) is expressed in cycles (f) is the cyclic frequency (cycle/s).
per second. The sine function is of innite length The frequency (f) is normally referred to as
extending along the time axis in both of the cyclic frequency, measured in cycles per second, to
negative and positive directions. Part of a sine differentiate it from the radian or angular frequency
function and its three characteristic parameters (x) which is measured in radiance per second.
are shown in Fig. 9.2. Frequency Units
The unit normally used in measuring frequency
is the Hertz (Hz), where 1 Hz is 1 cycle/s.
9.1.4 The Frequency Concept Another closely related measurementunit
applied in measurement of frequencies is the
Any periodic motion, as the rotationarm model Octave. Instead of quoting the frequency in its
shown in Fig. 9.1, is specied by two periodicity absolute units (Hertz), it is quoted in frequency
parameters: the frequency and the period of the ratios measured in Octaveunits. It is dened as
motion. Frequency is dened to be the number of the number of times (n) with which a frequency
revolutions (cycles) made in one second. The (f2) value gets doubled over a reference fre
length of the time interval (normally measured in quency (f1). A frequency ratio (f2/f1) is one
seconds or milliseconds) for one cycle, is called Octave if (f2/f1 = 2). In general, the number of
the period. The relation between frequency octaves (n) is related to the ratio (fn/f1) by:
(f) and period (s) is:
f n =f 1 2n
f 1=s n 1=log 2 logf n =f 1
n is number of Octaves
For the rotation arm, shown in Fig. 9.1, the
period (s) is the time interval during which the
slope slope
(in db/octave) (in db/octave)
filter frequency bandwidth
frequency
f1 f2 f3 f4
Fig. 9.3 Slopes of a fourcorner spectrum quoted in decibel (db) per octave units, where f2/f1 = f4/f3 = 2
9.1 The Seismic Processing Tools 201
s(t) s(t)
tp tp
t t
0 0
phase lag = tp phase lead, tp
s(t)
s(t)
tp tp
t t
0 0
Phase  90 + /4
lag
Zero 0 0
phase
Phase +  /4
lead
90
+
180 + /2
Outof 
phase 180  /2
202 9 The Seismic Processing Tools
s(t)
s(x)
a a
t x
Fig. 9.6 Denition of the period (s) and wavelength (k) of a sine function in the two domains, s(t) and s(x)
peaks, in a timedomain sine function. Likewise, simplest form, this type of function has the form
the distance separating two peaks in xdomain (Richardson 1953, p. 38):
sine function represents the wavelength (k) as
shown in Fig. 9.6. sx; t a sin 2pt=s x=k
In radians units, we have the temporal relation
(x = 2pft = 2p/s) and the corresponding spatial Using the relationships (x = 2pft = 2p/s) and
relation (k = 2pfx = 2p/k). Another important (k = 2pfx = 2p/k), the function can be written
relation is the one connecting the temporal as:
parameters with the spatial parameters, which is:
sx; t a sinxt k x
v k=s
where k is constant known as the wave number,
It should be noted that when frequency is v is propagation velocity, and x is the radian
quoted as (f) without the subscript (t), it is nor frequency measured in radians per second.
mally meant to be the temporal frequency (ft). It is important to note that the wave propa
gation velocity (v) is connected with both of the
temporal frequency (ft) and wavelength (k) by
9.1.7 Propagating Sine Wave the simple relationship:
v f tk
The rotating arm model (shown above), helped in
generating a one dimensional sine function that
or,
depends on one variable which is angle (h) or
time (t). The timedependent sine function v x=k
describes the case of stationary (nonpropagating)
periodic motion. However, we may have a Using a simple pictorial method, the velocity
twodimensional sine function s(x, t) describing formula (v = ft k) can be derived. Referring to
the case of a propagating sinusoidal wave. In its Fig. 9.7, the number of cycles (ft) of a
9.1 The Seismic Processing Tools 203
v = ft
propagating sine wave measured in 1 s is mul (iv) The integral of f(t)dt over a complete
tiplied by the cycle length (wavelength, k) gives period is convergent.
the total distance covered in that 1 s, which is the
It so happened that almost all geophysical
propagation velocity (v). Thus (v = ft k).
phenomena, including the seismic signal, obey
Dirichlet conditions and hence can be analyzed
by Fourier Theorem.
9.2 Fourier Analysis and Concept
of Spectra
9.2.1 Fourier Series
In 1807, Joseph Fourier (17681830) presented
his theorem (now identied by his name) which According to Fourier Theorem, any periodic
states that any function such as f(t), satisfying function, f(t) of period (s), satisfying Dirichlet
certain restrictions can be expressed as a sum of conditions, can be represented by the sum of
an innite number of sine waves. In 1809 sinusoidal functions of frequencies which are
Dirichlet formulated the mathematical restrictions multiples of the fundamental frequency of that
(normally referred to as Dirichlet conditions) function. Mathematically, the periodic function
under which the theorem is mathematically valid. f(t) is expressed by the following innite series
The restrictions are: (called Fourier series):
a0 =
an =
bn =
These three expressions are derived by mul of frequencies (xn) and amplitudes (an and bn).
tiplying both sides of the Fourier series in turn by The term (a0) represents the DC level of the
(1), (cos nxt), and (sin xt), and integrating with function f(t).
respect to (t) over the period length (s). The
Odd and Even Functions
derivation is based on use of the orthogonality
For an odd function fo(t) in which fo(t) = fo(t)
properties of the sine and cosine functions, which
are (for m and n being integers):
The frequencies (xn) that are multiples of the the coefcients (an) become all equal to zero,
fundamental frequency (x), are called harmonic since in this case:
an = = +
= +
=0
frequencies or just harmonics. Thus a periodic Similarly, it can be shown that, for an even
function f(t) is made up of the sum of an innite function fe(t) in which fe(t) = fe(t), the coef
number of harmonics (sine and cosine functions) cients (bn) become equal to zero for all (n).
9.2 Fourier Analysis and Concept of Spectra 205
Thus, when the function f(t) is odd, the approximation for the continuous parts of the
Fourier series will consist of sine terms only and function. This kind of distortions (synthesized
when it is an even function, the series will consist function overshooting and oscillationbehavior)
of cosine terms only. is normally referred to as Gibbs Phenomenon
(Fig. 9.8).
Application of the Fourier series analysis of
9.2.2 Gibbs Phenomenon periodic functions is found in many standard
texts on the subject, see for example (Alsadi
As we have stated above, a periodic signal (pe 1980, p. 111).
riod, s), satisfying other Dirichlet conditions, can
be expressed as a sum of sinusoidal components
of frequencies; n/s, (n = 0, 1, 2, 3, ). The 9.2.3 Fourier Transform
zerofrequency component (n = 0) represents the
DC term. This implies that if we have a signal, As it is presented above, Fourier series is
f(t) dened over the time interval (t = 0 to restricted to periodic functions, but theory can be
t = T), then, according to Fourier Theorem, the extended to cover nonperiodic functions. Func
function within this interval can be expressed as tions of nite length are called transient func
the sum of the frequency components (n/T) tions. With special integrals, it is possible to
where (n = 0, 1, 2, 3, , ). transform a transient function f(t) to another form
Fourier Theorem states that the signal f(t) will in which it becomes function of frequency F(x),
be recovered exactly from the Fourier series only with no loss of information. The process of
when an innite number of terms are included in converting f(t) into F(x) is called (Fourier
the summation, that is when the integer (n) runs transform) and the reverse process, F(x) into f(t),
from (n = 1) to (n = ). In practice, only a is called (inverse Fourier transform).
nite number (n = N, say) of terms is used in the The transient function can be considered as a
summation. As a result of this truncation of the periodic function of innitely long period. The
series (i.e. when n is nite number), distortion process of transforming a transient function from
shall occur in the recovered (synthesized) signal. its time domain f(t) to the frequency domain
In this case, the sum will shoot beyond the value function F(x) and vice versa, are done through a
of f(t) in the neighborhood of the discontinuities set of integral equations called (Fourier Integral
found in f(t). The overshoot oscillates about the equations) which are more commonly known as
function value with decreasing amplitude as we Fourier Transform equations. The two equations
move away from the discontinuity. Increasing (Transform and Inverse Fourier transform equa
the number of terms leads to a better tions) are Bth (1974, pp. 3337):
f(t)
f(t)
t t
(fewterms) approximation (moreterms)
Fig. 9.8 Gibbs Phenomenon computed for a rectangular pulse for few and moreterms truncation of Fourier series
206 9 The Seismic Processing Tools
(Fourier Transform)
In general, the function F(x) is complex, similar to Light spectrum in which the Light
consisting of a real part a(x) and imaginary part frequencycomponents appear when analyzed as
b(x). Thus, F(x) can be expressed as: it passes through a transparent prism. In the
The two functions, a(x) and b(x), known mathematical sense, F(x) is complex function
respectively as cosine and sine transforms, are consisting of the modulus; jFxj called (am
dened as: plitude spectrum) and the argument (x), called
or.
F() = A()ei()
where,
Z1 p
jFxj ax2 bx2
bx f t sin xt dt
x tan1 bx=ax
1
represents the Fourier spectrum of the analyzed phasespectra. Through another form of integra
function. tion (Inverse Fourier transform), it is possible to
recover the original timedomain function from
9.3.1 The Line Spectrum the transformed frequencydomain function. For
complete recovery, this process requires both of
Based on the nature of the input signal (periodic or the amplitude spectrum and phase spectrum. The
non periodic), the computed spectrum is obtained concept of frequency spectra (amplitude and
as discrete points or as continuous curve. Fre phasespectra) computed by Fourier transform
quency analysis of a periodic function involves equations are illustrated pictorially in Fig. 9.10.
determination of the Fourier series coefcients of In summary, it can be stated that
the input function. In this case the output Fourier series analysis of periodic functions
frequencydomain function is plotted as discrete yields line spectra and Fourier transform analysis
lines along the frequency axis forming what is of nonperiodic functions yields continuous
commonly referred to as the (line spectrum). spectra.
The line spectrum S1(x) of a sine function
s1(t) = a sinx1t, is represented by one line of
height proportional to the amplitude (a) of the 9.3.3 The Fourier Power Spectrum
sine function, located at frequency (x1). For an
another sine function, s2(t) = a sinx2t of fre The total energy (E) of a real function f(t), such
quency (x2), will, likewise, be of one line placed as a seismic signal, is generally taken to be
at frequency (x2). Further, the sum of the two proportional to the integral of the square of its
sine functions s1(t) and s2(t) will have a line amplitude. That is:
spectrum made up of two lines placed at (x1) and
Z1
(x2). The principle is shown in Fig. 9.9.
E ff tg2 dt
1
9.3.2 The Continuous Spectrum
It can be shown (Bth 1974, p. 82) that this
Analysis of a nonperiodic signal, by the Fourier expression is related to the power spectrum as
transform, gives continuous amplitude and follows:
s2(t)
S2( )
t
2
s1(t) S1( )
+ +
s2(t) S2( )
t
1 2
208 9 The Seismic Processing Tools
a1
a2 A()
a3 a0 a1 a2 a3 a4 a5
1 2 3 5
a4
() phase spectrum, ()
a5
Fourier Pair
f(t) F()
p(t) 2a
2a
t
a 0 +a 2 /a  /a 0 /a 2 /a
Fig. 9.12 The window pair of the rectangularpulse, boxcar time function, pa(t) and its spectrum, sinctype function,
P(x) = (2sin ax)/x
210 9 The Seismic Processing Tools
In the geophysical literature, this function is pulse rectangle gets narrower, the corresponding
known as the (sinc) function. The function (sinc amplitude spectrum gets wider, and in the lim
x) is dened as (sinx)/x or (sinpx)/px (Sheriff iting state, the pulse becomes spikefunction
2002, p. 320). Fourier transform P(x) of the (impulse function) with innitely wide spectrum
rectangular function (normally called boxcar (Fig. 9.13).
function, p(t) is the sinc function, 2a sinc (ax/p). Another important conclusion can be drawn
from this behavior, and that is the spike time
function contains innite number of equal
9.3.6 Spectrum of the Spike Pulse amplitude frequency components. For this rea
son the spectrum of a spike signal is often
Mathematical analysis proved that there is an described as being white spectrum.
inverse relationship between the width of the
timedomain function (as a seismic pulse) and its
frequency domain amplitude spectrum. Thus, as 9.3.7 The DiracDelta Function
the pulse gets narrower, the corresponding
frequencydomain spectrum becomes wider, and The zerowidth pulse (impulse) is called (Dirac
vice versa. This behavior can be readily seen in delta function) and given the symbol d(t). This
case of the rectangular pulsespectrum Fourier special function is dened as being a compressed
pair. The main lobe of its spectrum extends from rectangular pulse dened to be of width,
(x = p/a) to (x = +p/a). This clearly indicates approaching zero and having a unit area (Bth
that the spectrum has inverse relation to the 1974, p. 52). It is considered as being even
original pulse width (2a). In fact, this is one of function and hence it has a real spectrum func
the important properties that hold for all Fourier tion which is constant at the value of unity for all
pairs of pulseshaped functions. frequencies. This means that the Delta function
According to this principle (called the can be synthesized by superposing an innite
reciprocity property) it is readily seen that as the number of sine functions which are in phase at
f(t) F()
t
0 0
f(t)
F()
)
0 t 0
9.3 Concept of the Frequency Spectrum 211
+
+
t=0
one point (at t = 0) where the components will The signal length (T, say) denes the lowest, or
add up constructively, and destructively else fundamental frequency (fL = 1/T) in the spectrum.
where as shown in Fig. 9.14. Also, it denes the frequency increment, f (= 1/T)
Dirac delta function, d(t) is not a proper of the spectrum computation. The digitization
mathematical function. It is usually considered as interval, or sampling period (t), on the other hand,
a mathematical concept which possesses its own sets the upper frequency limit of the spectrum. In
mathematical properties. It is dened as: fact, the upper limit is the Nyquist frequency (fN)
which is equal to (1/2t). With consideration of
Z1 these limits, spectrum of an observed signal, of
dt 0 for t 6 0 and dtdt 1 nite length (T) and digitized at (t) interval, is
1 usually computed for the frequencies:
The delta function is even, d(t) = d(t), and it 1=T; 2=T; 3=T; . . .; 1=2Dt
forms the Fourier pair (d(t) $ 1). Further, we have
another important property which states that its Thus, a digital signal made up of (N + 1)
convolution with a function leaves the convolved samples, will be of length (T) given by
function unchanged, that is: d(t) * f(t) = f(t). (T = Nt). By substituting (T/N) for (t), com
putation frequencies will be:
Finally, it should be emphasized that the length These are expressed in the frequency domain by
(T), controls the (resolution power) of the com the amplitude and phasespectra as shown in
puted spectrum, since the frequency interval f Fig. 9.15.
(= 1/T) is inversely proportional to the length (T). In Fig. 9.15, signal1 has negative timephase
(positive phaseangle), while signal2 is of
zerophase. The rest of frequency components
9.4 The Phase Spectrum are of negative phase angles. These amplitude
and phase characteristics are shown in the cor
From the previous paragraph, we learned that the responding frequency domain spectra.
frequency spectrum F(x), of any real time
function is generally a complex function, con
sisting of the modulus; jFxj called (amplitude 9.4.1 The ZeroPhase Spectrum
spectrum) and the argument (x), called (phase
spectrum). The importance of the phase charac If all the frequency components are of zero phase
teristics of a seismic signal is that it has direct (symmetrical about time zero), their sum will
influence on the signal shape. Signal amplitude, give a zerophase wavelet, symmetrical about
on the other hand has a direct effect on the signal zero time. The phase spectrum, in this case, is a
energy. The phase of a frequency component is straight line drawn along the frequency axis at
measured by the time shift of a peak of a cycle zero phasevalue. Amplitude and the phase
with respect to time zero, considered to be the spectrum for this case are shown in Fig. 9.16.
start time. Thus, the start time is the time relative Note that the amplitude spectra in the two
to which all phases of the frequency components Figs. 9.15 and 9.16 are the same, but their
are referred to. In general, the phase value may respective time domain signalshapes are differ
be negative, zero or positive. ent. This difference in shape is due only to their
Seismic signals, like seismic reflection wave different phase spectra. Another useful note is
lets, are all of the type of signals which are that the recorded seismic reflection wavelet is
existing only in the positive side of the origin onesided signal. It cannot be zerophase since,
time. Such functions (called onesided functions) by denition, it has no energy before time zero as
consist of the sum of frequency components of onesided signal demands. However, conversion
different amplitudes and different phase shifts. into zerophase wavelet can be done by
amplitude spectrum
1 t
2
1 2 3 4
3
+
amplitude spectrum
1 t
2
1 2 3 4
3
+
4

zerophase spectrum
sum=
t=0
Fig. 9.16 Set of frequency components (sine functions) of a zerophase signal, with its amplitude and phase spectra
application of special computer programs. This frequency. In this case, each frequency compo
can be done in processing stage to convert the nent will be shifted by the same time shift
onesided wavelet to a zerophase wavelet of (Sheriff and Geldart 1995, p. 533). The sum, in
symmetrical shape. This is normally asked for by this case, will be like the zero phase spectrum
interpreters for more clear denition of reflection (symmetrical wavelet), but the symmetry will be
events and easier to follow in the process of about a time later than zerotime. The phase
interpretation especially for structural purposes. spectrum will be a straight line inclined to the
frequency axis (Fig. 9.17).
A closely associated case is the linear phase The constant phase spectrum is a special case of
spectrum, in which the phase is linear function of linear phase spectrum. The shift is not function of
amplitude spectrum
1 t
2 1 2 3 4
+
3

sum = linearphase spectrum
t=0
Fig. 9.17 Set of frequency components (sine functions) of a linearphase signal, with amplitude and phase spectra
214 9 The Seismic Processing Tools
frequency but having a constant value. Thus spectrum in this case is straight line parallel to
when the phase value is constant at zerovalue, the frequency axis and located at (p) below (or
we get a zerophase spectrum as shown in above) the frequency axis. The (p) or (+p)
Fig. 9.16. If the constant shift is a nonzero phase of a component implies that it has a trough
value, the phase spectrum is also a straight line, at time zero. The wavelet resulting from sum
but it is parallel to the frequency axis. For ming such components will be symmetrical and
example if the constant phase shift is at laterthan will have a trough at time zero as shown in
the time zero by quarter a period (s/4, say), Fig. 9.19.
which is equivalent to (p/2), the phase spectrum
will be a straight line located at (p/2) below the
frequency axis. In this case, the frequency com 9.5 Spectra of Observational Data
ponents will all have the constant phase of (p/2)
and their sum will give an invertedsymmetry As far as Fourier analysis is concerned, an
wavelet that has its main peak immediately fol observational function differs from an analytic
lowing the time zero (Fig. 9.18). function in two main aspects. The observed
If the phases of the frequency components are function is not innite and not continuous as an
all equal to (+p/2) instead of (p/2), the phase analytic function would be. Thus, a signal,
spectrum will be a straight line at (p/2) above the (normally, time function), inputted to the trans
frequency axis. In this case, all the components formation process, is nite in length and digital
will have troughs immediately following the time in form. Observational functions, which are input
zero, and the wavelet resulting from summing the for spectral analysis, are usually not continuous
components will, likewise, have a trough fol and not innite in length as Fourier Transform
lowing time zero. integral demands. For this reason, observational
The last interesting constantphase case is a spectra suffer from the two types of distortion
wavelet having all its component frequencies of which are resulting from signaltruncation and
equal phases, at the constant phase of (p), signaldigitization. These effects shall be dis
which is the same as (+p). The constantphase cussed as follows:
amplitude spectrum
1 t
2 1 2 3 4
3
+
4

constantphase spectrum
sum =
t=0
Fig. 9.18 Set of frequency components (sine functions) of a constantphase (at phase = p/2), with amplitude and
phase spectra
9.5 Spectra of Observational Data 215
amplitude spectrum
1 t
2
1 2 3 4
3
+
4

sum =
constantphase spectrum
t=0
Fig. 9.19 Set of frequency components (sine functions) of a constantphase (at phase = p, or at phase = +p), with
amplitude and phase spectra
This shows that signal truncation incurs a 9.5.2 The Rectangular Window
smoothing effect due to convolution of the true (BoxCar)
spectrum F(x) with the spectrum of the window
function, W(x). In addition of the smoothing In order to get the least possible spectrum dis
effect side lobes are also created. So, the spectrum tortion, the time interval of the window function
of a truncated signal is distorted, and the nature of must be as long and smooth (free of sharp cor
distortion is the smoothing effect and develop ners) as it can be (Bth 1974, pp. 155171).
ment of side lobes as shown by Fig. 9.20. Many types of window functions have been
Since an innitely long rectangular pulse has developed for application in this eld (8 window
a spectrum approaching an impulse function, it types were presented by Sheriff 2002, p. 397).
timedomain frequencydomain
truncation process convolution process
f(t) . w(t) F( ) * W( )
 R( ) * F( )
Fig. 9.20 Effect of varying the truncation window width (T) on smoothing of the spectrum F(x) of the truncated signal
f(t). Smoothing is more severe with shorter length, T
9.5 Spectra of Observational Data 217
t  /T /T
T 0 T  0
t
T T  2 /T 2 /T
Fig. 9.21 The two time windows in common use in seismic data processing and their spectra. a The triangular
timefunction, or Boxcar window and b the triangular timefunction, or Bartlett window
218 9 The Seismic Processing Tools
This denition can be represented in the fol where, () represents correlation process, (s) is
lowing symbolical way: called the time lag.
f1 ( i ) = 2 2 1 1
f2 ( i ) = 2 1 0 1
2 1 0 1
2 1 0 1
2 1 0 1
2 1 0 1
2 1 0 1
2 1 0 1
C12 ( j ) = 2 2 1 7 5 1 2
f1 f2 C12
f1 ( i ) = 2 2 1 1
2 2 1 1
2 2 1 1
2 2 1 1
2 2 1 1
2 2 1 1
2 2 1 1
C11 ( j ) = 2 0 5 10 5 0 2
f1 f1 C11
In practice, it is a shift, multiply, and sum maxima at a period equal to the periodicity of the
process as it is done in computing the hidden periodic signal if it is existing in the orig
crosscorrelation function (Fig. 9.23). inal function. Most intensive use of the autocorre
The function C11(s) is even function, lation functions is in computing power spectra and
C11(s) = C11(s), that is, symmetrical about the in the deconvolution computations.
point (s = 0).
It is useful to note here that the autocorrelation function in the time domain is transformed into
power spectrum in the frequency domain. It is also important to note that the phase spectrum, 11(x),
is equal to zero. This provides an alternative method which allows computing power spectra from the
autocorrelation function instead of squaring the amplitude spectrum.
9.7 Convolution
Like correlation functions, Convolution is a mathematical process taking place between two functions
to output a new function the convolution output. Convolution of the two functions; f(t) and h(t) is
dened by the integral:
yj = fi * hi = fi . hji
Computation of the convolution process is done in the same way as in the correlation process
(described above) except for the need to reverse one of the two functions involved in the process.
Here is an example to clarify the process of convolving fi with hi where fi = { 2 1 3 0 1} and
hi = { 3 1 2}. The convolution process starts with reversing one of the two functions and then
continues with shifting, multiplication, and summing processes, shown as follows:
f( i ) = 2 1 3 0 1
h(i)= 2 1 3
2 1 3
2 1 3
2 1 3
2 1 3
2 1 3
2 1 3
y (i) = 6 5 14 5 3 1 2
There are several other methods to compute convolution. These methods involve use of: special
tables, special matrix, and by ztransform. Methods of computation and mathematical properties of
the convolution process are summarized in Alsadi (1980, p. 8890).
The process involved in convolution (shiftmultiplyandsum process) is shown graphically in
Fig. 9.24, in which the function f(t) is convolved with the function h(t).
9.7 Convolution 221
t t
0 0
f(t) F()
h(t) H()
Much more than documents.
Discover everything Scribd has to offer, including books and audiobooks from major publishers.
Cancel anytime.